Description

PATENT SEARCHING METHOD AND PATENT SEARCH

SYSTEM

Technical Field

[1] The present invention relates, in general, to a method and system for providing a patent search service, which automatically includes and then searches for hierarchically dependent patent classification terms, and, more particularly, to a method and system for providing a patent search service, which automatically includes and then searches for patent classification terms that share the lowest classification term representation format from the point of view of the representation of patent classification terms and are hierarchically dependent from the point of view of content. Background Art

[2] Almost all patent documents are each assigned at least one classification term of the

International Patent Classification (IPC), which indicates the technology included in the patent document using a single symbol. The IPC was established under the Strasbourg agreement. In the IPC, almost all patent documents are each assigned at least one IPC classification term. The IPC has a classification term system that represents classification terms using sections, classes, subclasses, groups, and subgroups in a hierarchical manner.

[3] Meanwhile, the patent offices of several countries have respective unique patent classification systems. Representative examples thereof include the U.S. Patent Classification (USPC) used by the United States Patent and Trademark Office (USPTO), the European CLAssification (ECLA) used by the European Patent Office (EPO), and the File Indexes (FI) and the F-terms (FT) used by the Japanese Patent Office (JPO).

[4] The USPC uses a class and subclass term system, the Japanese FI has a structure in which the IPC has been extended vertically, and the Japanese FT has a higher subject class called a 'theme' and an F-term system relevant to the structure below the theme.

[5] In the patent classification term systems, each classification term is associated with a part indicating the meaning of the relevant patent classification term, that is, "title," in a 1:1 correspondence. For example, IPC HOlS 3/00 is associated with the title "laser", and HOlS 3/14 is associated with the title "• characterized by the material used as the active medium."

[6] The above-described hierarchical patent classification term and title systems share a dot structure as a common characteristic therebetween. Taking the IPC as an example, all patent classification terms relevant to subgroups have the structure "subclass'V'number/number," and most titles relevant to these patent classification

terms are each combined with one or more dots. All titles except for titles relevant to classification terms including 'number/00' are each combined with one or more dots, and 'number/00' is generally a symbol relevant to a group. When the number of dots combined with a title is smaller, the title pertains to a higher technical class than adjacent titles. In contrast, when the number of dots is larger, the title pertains to a relatively lower/detailed technical class compared to adjacent titles.

[7] An example of a hierarchical patent classification term system including a dot structure is shown below.

[8] section: H ELECTRICITY

[9] class: HOl BASIC ELECTRICAL ELEMENTS

[10] subclass: HOlF MAGNETS

[11] main group: HOlF 1/00 Magnets or electromagnets characterized by the magnetic material

[12] 1-dot subgroup: 1/01 • Of inorganic materials

[13] 2-dot subgroup: 1/03 •• Characterized by their coercivity

[14] 3-dot subgroup: 1/032 ••• Of hard-magnetic materials

[15] 4-dot subgroup: 1/04 •••• Metals or alloys

[16] 5-dot subgroup: 1/047 ##### Alloys characterized by their composition

[17] 6-dot subgroup: 1/053 ###### Containing rare earth metals

[18] In the above example, it can be seen that there is a hierarchical structure even between subgroups having the same patent classification term representation format from the point of view of technical content.

[19] However, since the classification term representation format is used at a subgroup level, the following problems occur when patent searches are performed using patent classification terms.

[20] First, when a search is generally performed using a patent classification term, a problem arises in that a search engine searches only for documents consistent with the entered patent classification term, without considering the hierarchical structure. Referring to the above example, when HOlF 1/032 is entered, search results consistent with a searcher's intention must include all documents that are assigned patent classification terms relevant to all lower classes conceptually included in HOlF 1/032, such as HOlF 1/04, HOlF 1/047 and HOlF 1/053, as well as HOlF 1/032.

[21] Second, when search results obtained using a patent classification term are analyzed, analysis results that include a logical error may be provided because the hierarchical structure of patent classification terms below the subgroup is not considered. That is, when the number of patent applications is analyzed, the number of documents that fall under HOlF 1/032 is always larger than or equal to the number of documents that fall under HOlF 1/04, regardless of the representation of patent classification terms

indicated in patent documents. That is, it is obvious that a patent document assigned HOlF 1/04 falls under HOlF 1/032. Since this document also falls under HOlF 1/04, which is one of the lower classes of HOlF 1/032, this document is assigned HOlF 1/04, rather than HOlF 1/032.

[22] Accordingly, in a patent search or patent analysis, it is necessary to include the dot structure information of titles relevant to patent classification terms, to automatically extract a lower patent classification term relevant to a patent classification term when the patent classification term is included in a search query entered by a searcher, and to provide search results in which the lower patent classification term has been considered.

Disclosure of Invention Technical Problem

[23] The first object of the present invention is to provide a method and system for providing a patent search service, which, when a patent classification term is included in a search query, can include and then search for lower patent classification terms that are hierarchically dependent on the patent classification term.

[24] The second object of the present invention is to provide a method of supporting a range search that is introduced to prevent a reduction in search speed that occurs in the case in which, when a patent classification term is included in a search query, lower patent classification terms, which are hierarchically dependent on the patent classification term, are included and then searched for, and a system for implementing the method. Technical Solution

[25] In order to accomplish the above objects, the present invention provides a method of providing a search service, in which a patent search service provision system for receiving a search query from a user, searching one or more of a search index and a patent information DB established for patent information, and providing search results provides the search service, the method comprising the steps of the patent search service provision system receiving a search request in which a patent classification term included in a preset patent classification term system is included in a search query; the patent search service provision system obtaining a next sibling patent classification term of the patent classification term, included in the search query, from the patent classification term system; the patent search service provision system creating a modified search query including the patent classification term and the next sibling patent classification term; and the patent search service provision system performing a search using the modified search query through a search engine; wherein the modified search query includes patent classification term range information that is created using

the patent classification term and the next sibling patent classification term; wherein the patent classification term system is any one of an International Patent Classification (IPC), a U.S. Patent Classification (USPC), a Japanese File Index (FI), a Japanese F- term (FT), and an European patent CLAssification (ECLA), and search results for hierarchically dependent lower patent classification terms are automatically included through a range search; and wherein the obtaining the next sibling next sibling patent classification term of the patent classification term included in the search query is obtaining the next sibling next sibling patent classification term in such a way that the patent search service provision system queries a hierarchical patent classification term DB included in the patent search service provision system.

[26] Preferably, the patent classification term included in the search query is a dot level- patent classification term; the lower patent classification terms are all patent classification terms below the patent classification term included in the search query; and the dot level-patent classification term is title information, that is, a description relevant to the patent classification term in the patent classification term system, in which one or more dots are included.

[27] Preferably, the dot level-patent classification term is, when the patent classification term system is the IPC, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subgroups, in which one or more dots are included; when the patent classification term system is the Japanese FI, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subgroups, in which one or more dots are included; when the patent classification term system is the Japanese FT, title information, that is, a description relevant to the patent classification term among patent classification terms below theme codes, in which one or more dots are included; and when the patent classification term system is the ECLA, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subgroups, in which one or more dots are included.

[28] Preferably, the step of the patent search service provision system creating the modified search query comprises the steps of creating the patent classification term range information about a range greater than or equal to the patent classification term included in the search query and less than the next sibling patent classification term with reference to a hierarchical patent classification term system DB in which the patent classification terms are sequentially arranged in dictionary order; and modifying the search query by considering the created patent classification term range information.

[29] Preferably, the step of the patent search service provision system searching for the next sibling next sibling patent classification term of the patent classification term

included in the search query is performed using any one of a method of querying the hierarchical patent classification term DB, in which the hierarchically organized patent classification term system has been considered, about the patent classification term included in the search query, and a method of searching a tree structure of all the patent classification terms using a depth-first search technique with reference to the tree structure of all the patent classification terms representing a tree structure of the patent classification term system.

[30] Preferably, the hierarchical patent classification term DB is created using any one of a method of creating the DB while considering dot structure information included in title information, that is, descriptions relevant to the patent classification terms in the patent classification term system, and a method of creating the DB using data indicating relationships between the given patent classification term and its immediately higher patent classification terms.

[31] Preferably, the tree structure of all the patent classification terms representing the tree structure of the patent classification term system is created using any one of a method of creating the DB while considering dot structure information included in title information, that is, descriptions relevant to the patent classification terms in the patent classification term system, and a method of creating the DB using data containing relationships between the given patent classification term and its immediately higher patent classification terms.

[32] In order to accomplish the above objects, the present invention provides a method of providing a search service, in which a patent search service provision system for performing a search for an obtained search query for patent information and outputting search results provides the search service, the method comprising the steps of the patent search service provision system receiving a search request in which a patent classification term included in a preset patent classification term system is included in a search query; the patent search service provision system performing any one of the step a) of the patent search service provision system obtaining a next sibling next sibling patent classification term of the patent classification term, included in the search query, from the patent classification term system, and obtaining a modified patent classification term and a next sibling modified patent classification term respectively relevant to the patent classification term and the next sibling patent classification term, and the step b) of the patent search service provision system obtaining a modified patent classification term relevant to the patent classification term and a next sibling modified patent classification term of the modified patent classification term; the patent search service provision system creating a modified search query including the modified patent classification term and the next sibling modified patent classification term; and the patent search service provision system performing a search

using the modified search query through a search engine; wherein the modified search query includes modified patent classification term range information that is created using the modified patent classification term and the next sibling modified patent classification term; wherein the patent classification term system is any one of an IPC, an USPC, a Japanese FI, a Japanese FT and an ECLA; wherein the obtaining the next sibling next sibling patent classification term of the patent classification term included in the search query is obtaining the next sibling next sibling patent classification term in such a way that the patent search service provision system queries a hierarchical patent classification term DB included in the patent search service provision system; and wherein the obtaining the modified patent classification term and the next sibling modified patent classification term is obtaining the modified patent classification term and the next sibling modified patent classification term in such a way that the patent search service provision system queries the hierarchical modified patent classification term DB included in the patent search service provision system.

[33] Preferably, the patent classification term included in the search query is a dot level- patent classification term; the lower patent classification terms are all patent classification terms below the patent classification term included in the search query; and the dot level-patent classification term is title information, that is, a description relevant to the patent classification term in the patent classification term system, in which one or more dots are included.

[34] Preferably, the dot level-patent classification term is, when the patent classification term system is the IPC, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subgroups, in which one or more dots are included; when the patent classification term system is the USPC, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subclasses, in which one or more dots are included; when the patent classification term system is the Japanese FI, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subgroups, in which one or more dots are included; when the patent classification term system is the Japanese FT, title information, that is, a description relevant to the patent classification term among patent classification terms below theme codes, in which one or more dots are included; and, when the patent classification term system is the ECLA, title information, that is, a description relevant to the patent classification term among patent classification terms relevant to or below subgroups, in which one or more dots are included.

[35] Preferably, the step of the patent search service provision system creating the modified search query comprises the steps of creating modified patent classification term range information about a range greater than or equal to the modified patent clas-

sification term in the search query and less than the next sibling modified patent classification term with reference to a hierarchical modified patent classification term system DB in which modified patent classification terms relevant to the patent classification terms are sequentially arranged in dictionary order; and modifying the search query by considering the created modified patent classification term range information.

[36] Preferably, the step of the patent search service provision system searching for the next sibling next sibling patent classification term of the patent classification term included in the search query is performed using any one of a method of querying the modified patent classification term DB, in which the hierarchically organized modified patent classification term system has been considered, about the modified patent classification term included in the search query, and a method of searching a tree structure of all the modified patent classification terms using a depth-first search technique with reference to the tree structure of all the modified patent classification terms representing a tree structure of the modified patent classification term system.

[37] Preferably, the modified patent classification term DB is created using a method of arranging all the patent classification terms of the patent classification term system having a hierarchical structure in a tree structure created by considering dot structure information included in title information, that is, descriptions relevant to the patent classification terms; and associating the modified patent classification terms with patent classification terms at respective nodes of the tree structure in a 1 : 1 correspondence while considering a sequence that is determined by searching the tree structure using a depth-first search technique.

[38] Preferably, the modified patent classification term system, created by associating the modified patent classification terms with patent classification terms at respective nodes of the tree structure in a 1 : 1 correspondence while considering a sequence that is determined by searching the tree structure using a depth-first search technique, is configured such that the sequence obtained through the search is consistent with a corresponding sequence obtained when the modified patent classification terms are arranged in dictionary order.

[39] Preferably, the search engine includes an index; the index further includes information about modified patent classification terms capable of accommodating the modified search query; the modified patent classification terms are arranged in the index in dictionary order because they can be arranged in dictionary order; and the modified search query includes the modified patent classification term.

[40] In order to accomplish the above objects, the present invention provides a method of providing a search service, in which a patent search service provision system for performing a search for an obtained search query for patent information and outputting search results provides the search service, the method comprising the steps of the

patent search service provision system receiving a search request in which a patent classification term included in a preset patent classification term system is included in a search query; the patent search service provision system creating a modified search query so as to produce search results related to lower patent classification terms of a patent classification term, included in the search query, in the patent classification term system; and the patent search service provision system performing a search using the modified search query through a search engine; wherein the search engine includes an index; wherein the index further includes information about modified patent classification terms capable of accommodating the modified search query; wherein the modified patent classification terms are arranged in the index in dictionary order because they can be arranged in dictionary order; wherein the modified search query includes information about a next sibling modified patent classification term of the modified patent classification term; and wherein the patent classification term system is any one of an IPC, an USPC, a Japanese FI, a Japanese FT and an ECLA.

[41] Preferably, the modified patent classification terms are associated with the patent classification terms in a 1 : 1 correspondence, and have a hierarchical structure identical to the hierarchical structure of the patent classification terms.

[42] Preferably, the step of the patent search service provision system creating the modified search query is performed using any one of a method comprising the steps of the patent search service provision system searching for a next sibling next sibling patent classification term relevant to the patent classification term included in the search query; the patent search service provision system obtaining a modified patent classification term relevant to the patent classification term included in the search query and a modified patent classification term relevant to the next sibling patent classification term; the patent search service provision system creating modified patent classification term range information about a range greater than or equal to the modified patent classification term relevant to the patent classification term included in the search query and less than the modified patent classification term relevant to the next sibling; and the patent search service provision system modifying the search query by considering the created modified patent classification term range information; and a method comprising the steps of the patent search service provision system searching for a modified patent classification term relevant to the patent classification term included in the search query; the patent search service provision system searching for a next sibling next sibling modified patent classification term of a modified patent classification term relevant to the patent classification term included in the search query; the patent search service provision system creating modified patent classification term range information about a range greater than or equal to the modified patent classification term relevant to the patent classification term included in the search query

and less than the next sibling modified patent classification term; and the patent search service provision system modifying the search query by considering the created modified patent classification term range information; wherein the obtaining the modified patent classification term and the next sibling modified patent classification term is obtaining them in such a way that the patent search service provision system querying a hierarchical modified patent classification term DB included in the patent search service provision system.

[43] Preferably, the step of the patent search service provision system searching for the next sibling next sibling patent classification term of the patent classification term included in the search query is performed using any one of a method of querying a hierarchical patent classification term DB created by considering the hierarchically organized patent classification term system about the patent classification term included in the search query and a method of referring to a tree structure of all the patent classification terms representing the tree structure of the patent classification term system; and the step of the patent search service provision system searching for a next sibling next sibling modified patent classification term of the modified patent classification term is performed using any one of a method of a modified patent classification term DB created by considering the hierarchically organized modified patent classification term system about the modified patent classification term and a method of referring to a tree structure of all the patent classification terms representing the tree structure of the modified patent classification term system.

[44] Preferably, the hierarchical patent classification term DB is created using any one of a method of creating the DB by considering dot structure information included in title information, that is, descriptions relevant to the patent classification terms in the patent classification term system, and a method of creating the DB using data indicating relationships between the given patent classification term and its immediately higher patent classification terms; and the modified patent classification term DB is formed in a structure identical to a structure of the hierarchical patent classification term DB, and is associated with the respective patent classification terms constituting contents of the hierarchical patent classification term DB in a 1 : 1 correspondence.

[45] Preferably, the tree structure of all the patent classification terms representing the tree structure of the patent classification term system is created using any one of a method of creating the DB by considering dot structure information included in title information, that is, descriptions relevant to the patent classification terms in the patent classification term system, and a method of creating the DB using data indicating relationships between the given patent classification term and its immediately higher patent classification terms; and the tree structure of all the modified patent classification term representing the tree structure of the modified patent classification term

system has a structure identical to the tree structure of all the patent classification terms representing the tree structure of the patent classification term system, and is associated with the respective patent classification terms constituting contents of the tree structure of all the patent classification terms representing the layer tree of the patent classification term system in a 1:1 correspondence.

[46] In order to accomplish the above objects, the present invention provides a search service provision system for receiving a search query from a user, searching one or more of a search index and a patent information DB established for patent information, and providing search results, the system comprising a modified patent classification term DB for storing modified patent classification terms modified to be associated with patent classification terms in at least one patent classification term system in a 1:1 correspondence based on information about parent-child relationships between the patent classification terms, having a hierarchical structure in the patent classification term system; a modified patent classification term DB creation module for creating the modified patent classification term DB; a lower patent classification term processing module for, when a patent classification term is included in the search query and the patent classification term is a patent classification term having one or more dots in title information relevant to the patent classification term in the patent classification term system, obtaining information about a lower patent classification term of the entered patent classification term or information about a lower modified patent classification term of the modified patent classification term; and a modified search query creation module for modifying the entered search query so that the search query includes information about the lower modified patent classification term, which is relevant to the lower patent classification term of the patent classification term having one or more dots and found by the lower patent classification term processing module; wherein the lower patent classification term processing module obtains the information about the lower patent classification term of the entered patent classification term by querying the hierarchical patent classification term DB included in the patent search service provision system; and wherein the lower patent classification term processing module obtains the information about the lower modified patent classification term of the modified patent classification term by querying the hierarchical modified patent classification term DB included in the patent search service provision system.

[47] Preferably, a modified patent classification term system including modified patent classification terms associated with patent classification terms in the patent classification term system having a hierarchical structure in a 1:1 correspondence has a hierarchical structure equivalent to the hierarchical structure of the patent classification term system.

[48] Preferably, the modified search query created by the modified search query creation

module, when the patent classification term information included in the entered search query is relevant to a classification term having one or more dots, includes information about the modified patent classification term relevant to the patent classification term having one or more dots and information about a next sibling modified patent classification term of the modified patent classification term relevant to the patent classification term in the modified patent classification term system.

[49] Preferably, the patent classification term system is any one of an IPC, an USPC, a

Japanese FI, a Japanese FT and an ECLA.

[50] Preferably, the system further includes a search engine for performing a search using the search query, the search engine includes an index; the index further includes information about modified patent classification terms capable of accommodating the modified search query; and the modified patent classification terms are arranged in the index in dictionary order because they can be arranged in dictionary order.

[51] Preferably, the modified patent classification term DB creation module creates the modified patent classification term DB using a method comprising arranging all the patent classification terms of the patent classification term system having a hierarchical structure in a tree structure; and associating the modified patent classification terms with patent classification terms at respective nodes of the tree structure in a 1 : 1 correspondence while considering a sequence that is determined by searching the tree structure using a depth-first search technique.

[52] Preferably, the modified patent classification term system, created by associating the modified patent classification terms with patent classification terms at respective nodes of the tree structure in a 1 : 1 correspondence while considering a sequence that is determined by searching the tree structure using a depth-first search technique, is configured such that the sequence of the searching is consistent with a corresponding sequence occurring when the modified patent classification terms are arranged in dictionary order.

[53] Preferably, the modified search query created by the modified search query creation module, when the patent classification term information included in the entered search query is relevant to a classification term having one or more dots, includes information about a range that is greater than or equal to the modified patent classification term, relevant to the patent classification term having one or more dots, and less than a next sibling modified patent classification term of the modified patent classification term in the modified patent classification term system.

Advantageous Effects

[54] The effects of the present invention are as follows:

[55] First, there may be hierarchical relationships between patent classification terms

having the same patent classification term representation format at a patent classifi cation term level at which title information, explaining patent classification terms, has one or more dots. In this case, a typical search engine searches only for patent documents that match an entered patent classification term. In contrast, when the spirit of the present invention is used, not only patent documents matching the entered patent classification term but also patent documents matching lower patent classification terms can be searched for as search results if there are lower patent classification terms that fall under the patent classification term in a patent classification term system. Through this, the number of patent documents that are excluded because they logically fall under an entered patent classification term but have different patent classification terms can be minimized.

[56] Second, the spirit of the present invention can be applied to patent monitoring for performing searches on new patent documents or updated patent documents using a preset search query and patent analysis for statistically analyzing search results. Through this, patent documents that are excluded because they logically fall under an entered patent classification term but have different patent classification terms are included in the targets of monitoring or analysis. Furthermore, the accuracy of monitoring and the accuracy of a set of analysis targets are increased, so that accurate monitoring and the accuracy of the set of analysis targets enable accurate analysis.

[57] Third, the number of patent documents relevant to a specific patent classification term can be accurately counted using the spirit of the present invention. Accordingly, since various types of physical quantities (speed, acceleration, increase rate, occupation rate, concentration rate, and an Activity Index (AI) and various statistics (average and standard deviation), related to the numbers of documents, can be accurately measured, analysis can be accurately performed.

[58] Four, there is an advantage in that search speed is increased when the range search of the lower classification inclusion search scheme of the present invention is performed. Brief Description of the Drawings

[59] Fig. 1 is a diagram showing the construction of an embodiment of a system for providing a patent search service according to the present invention;

[60] Fig. 2 is a diagram showing the flow of the provision of a patent search service for automatically including hierarchically dependent patent classification terms and performing searches according to the present invention;

[61] Fig. 3 is a diagram showing the construction of another embodiment of a system for providing a patent search service according to the present invention;

[62] Fig. 4 is a diagram showing an embodiment of a method of creating a hierarchical

modified patent classification term DB according to the present invention;

[63] Fig. 5 is a diagram showing an embodiment of a method of constructing a tree structure using USPC patent classification terms;

[64] Fig. 6 is a diagram showing an embodiment of a method of assigning modified patent classification terms to the USPC patent classification terms of Fig. 5 in accordance with the tree structure;

[65] Fig. 7 is a diagram showing an embodiment of a tree structure formed of modified patent classification terms having the same structure as the tree structure formed of the USPC patent classification terms of Fig. 5;

[66] Fig. 8 is a diagram showing the possibility of constructing the tree structure of patent classification terms, such as that shown in Fig. 5, using the Index to U.S. Patent Classification File (a.k.a., Classification Index File);

[67] Fig. 9 is a diagram showing an embodiment of a method of searching for modified patent classification terms and performing a range search according to the present invention;

[68] Fig. 10 is a diagram showing another embodiment of a method of searching for modified patent classification terms and performing a range search according to the present invention;

[69] Fig. 11 is a diagram showing an embodiment of a system for performing searches using a DBMS according to the present invention; and

[70] Fig. 12 is a diagram showing another embodiment of a system for searching for modified patent classification terms using a DBMS according to the present invention. Mode for the Invention

[71] The present invention will be described in detail below with reference to the accompanying drawings.

[72] Fig. 1 is a diagram showing an embodiment of a system 100 for providing a patent search service capable of including hierarchically dependent lower patent classification terms. The patent search service provision system 100 of the present invention allows the access of a user's computer, connected over a wired/wireless network 400, in a network manner, receives a search query from the user, creates a modified search query by processing a patent classification term using a patent classification term processing engine 140 if the patent classification term is included in the search query, and produces search results in such a way that a search engine 110 queries a search index 120 using the modified search query. It is natural that a patent information DataBase (DB) capable of providing patent information corresponding to the patent search results should exist in the present system.

[73] The search engine 110 of the present invention indexes the content information of

obtained patent documents using a search indexer, and stores it in the form of a search index 120. Here, the content of patent documents may be indexed for respective fields. Furthermore, the search engine 110 extracts found patent documents, and the patent search service provision system may retrieve all patent information related to the bibliographic information, body text and/or drawings of the found patent documents from a patent information DB 150 and transmit it to the user computer 100.

[74] A typical patent document basically includes bibliographic information, body information in the form of text in which the details of patent technology are included, and drawing information. In some cases, a patent document may further include information about administrative measures taken for the patent document. The bibliographic information includes at least one patent classification term.

[75] Patent classification term information is associated with the title information of the above-described patent classification information in a 1 : 1 correspondence. In the case of the IPC, hierarchical information, which cannot be represented using patent classification terms, is typically represented using the numbers of dots at levels below the subgroup level. In the case of the IPC, at levels below the subgroup level, because it is difficult to detect hierarchical relationships using only patent classification term information, hierarchical position information is detected using one or more dots that are included in title information that corresponds to patent classification term information. That is, at levels below subgroup level, the hierarchical structure between patent classification terms at levels below the subgroup level can be found by determining dot structures that are included in titles. The hierarchical structure between patent classification terms has the form of a tree structure.

[76] Although, in the present specification, the tree structure between patent classification terms will be described using IPC 7 version main group H04B 7/00 (the title thereof is "Radio Transmission Systems") and patent classification terms falling under the main group as an embodiment, the method and the description are not limited to that embodiment, but are applied to the entire IPC in the same or an equivalent manner. Furthermore, the method and the description are applicable to other patent classification term systems (for example, the USPC, the FI, the FT and the ECLA) that include dot structures in titles as a way of representing hierarchical position, in the same or an equivalent manner.

[77] The main group H04B 7/00 and patent classification terms and the titles thereof falling under the main group have the following configuration based on IPC 7 version, as of May 5, 2006.

[78] H04B 7/00 Radio transmission systems, i.e. using radiation field

[79] H04B 7/005 • Control of transmission; Equalization

[80] H04B 7/01 • Reduction of phase shift

[81] H04B 7/015 • Reduction of echo effects [82] H04B 7/02 • Diversity systems [83] H04B 7/04 ## Using a plurality of spaced independent aerials [84] H04B 7/06 ••• At transmitting station [85] H04B 7/08 ••• At receiving station [86] H04B 7/10 •• Using a single aerial system characterized by its polarization or directional properties

[87] H04B 7/12 ## Frequency-diversity systems

[88] H04B 7/14 • Relay systems

[89] H04B 7/145 •• Passive relay systems

[90] H04B 7/15 Active relay systems

[91] H04B 7/155 ¹ Ground-based stations

[92] H04B 7/165 Employing angle modulation

[93] H04B 7/17 Employing pulse modulation

[94] H04B 7/185 Space-based or airborne stations

[95] H04B 7/19 Earth-synchronous stations

[96] H04B 7/195 Non-synchronous stations

[97] H04B 7/204 Multiple access

[98] H04B 7/208 Frequency-division multiple access

[99] H04B 7/212 ¹ Time-division multiple access

[100] H04B 7/216 ¹ Code-division or spread-spectrum multiple access

[101] H04B 7/22 t Scatter propagation systems

[102] H04B 7/24 # For communication between two or more posts

[103] H04B 7/26 •• At least one of which is mobile

[104] When the above-described configuration is arranged to more easily construct a tree hierarchical structure based on the dots, it can be represented using the following Table 1:

[105] Table 1

[106] It will be apparent that, when the content of patent documents is indexed, the indexing is performed for each of the patent offices of respective countries and patent document processing organizations (in particular, the EPO and the WIPO).

[107] The search engine 110 of the present invention receives a search query from the user of the search engine 110, performs a search, and produces search results.

[108] The patent classification term processing engine 140 of the present invention basically includes sub-modules for processing the following tasks.

[109] A first sub-module is a hierarchical patent classification term DB creation module

142, and creates a hierarchical patent classification term DB 130, including a tree structure between patent classification terms, for patent classification information data including patent classification term information and the content of technology corresponding to the patent classification term information. The hierarchical patent classification term DB 130 may be taken as an example of a hierarchical patent classification term DB 130 because it is created in consideration of the hierarchical structure between patent classification terms.

[110] The hierarchical patent classification term DB creation module 142 functions to associate one patent classification term, other than the highest patent classification term (for example, a patent classification term representative of a section in the IPC, and a patent classification term representative of a class in the USPC) with an immediately higher patent classification term under which the patent classification term falls (1:1), or functions to associate the one patent classification term with patent clas-

sification terms immediately below the classification term (l:n). Although both methods are possible from the point of view of computer science, the former method of creating a 1 : 1 relationship in which one patent classification term (child) is associated with a higher term (parent) is more reasonable. Through the relationship association, all patent classification terms establish relationships therebetween within a tree structure.

[I l l] The former method is a method of associating a child with a parent, while the latter method is a method of associating a parent with children that fall under the parent and storing the association. The two methods can be used in a process of searching for children that fall under parents. In the former method, when a parent is given, a child who points to the parent can be searched for. In contrast, in the latter method, since information about the child that falls under the parent is directly stored, a child can be searched for using the stored information. As a result, the former method is a method of storing a child-parent relationship in a 1 : 1 association based on a child, while the latter method is a method of directly storing information about immediately lower children that fall under a parent based on the parent.

[112] The method in which the hierarchical patent classification term DB creation module

142 creates the hierarchical patent classification term DB 130 will be described based on the IPC classification below.

[113] 1) The hierarchical patent classification term DB 130 is created according to the classification term system for a range from a section to a main group, and a reference term's parent information or a reference term's immediately lower child information is stored for a patent classification code of a reference term, patent classification codes immediately above the reference term, and patent classification codes immediately below the reference term.

[114] 2) For a main group patent classification term, all patent classification terms before the next main group patent classification term are searched for a patent classification term, the title of which has one dot. Since a main group patent classification term is generally configured such that a slash "/" is combined with OO', it is easy to search for the main group patent classification term. In the case of a child -> parent method (a method in which a child points to its sole parent), the patent classification term of a main group, which is its sole immediately higher parent, is stored for each patent classification term of a 1-dot subgroup, which has one dot. In the case of a parent -> child method (a method of storing all children under a reference term), the patent classification terms of a 1-dot subgroup, which fall under the main group and each have one dot, are stored in association with a patent classification term representing the main group.

[115] In this case, a main group's parent information or the main group's immediately

lower child information is stored for the main group itself or its immediately lower patent classification code.

[116] 3) All patent classification terms between a patent classification term having one dot and a subsequent patent classification term falling under the same main group and having one dot are searched for one or more patent classification codes each having two dots in the titles thereof. If there are not other patent classification terms falling under the same main group and having one dot in the title thereof (if a main group has a single subgroup, or a relevant subgroup is the last one of relevant subgroups), all patent classification terms between a patent classification term having one dot in the title thereof and a patent classification term representing a subsequent main group are searched for one or more patent classification codes each having two dots in the titles thereof. A method of storing a reference term's parent and children is the same as that of the above paragraph 2).

[117] 4) The patent classification codes each having two dots in the titles thereof are searched for one or more patent classification codes each having three dots in the titles thereof in the same manner as in the above paragraph 3), and the subsequent processing is performed in the same manner as in the paragraph 3).

[118] Another method in which the hierarchical patent classification term DB creation module 142 creates the hierarchical patent classification term DB 130 will be described below.

[119] The method is a method of searching for a reference term's immediately higher parent while moving upwards from the lowest IPC term in a table in which IPC terms are sequentially arranged (according to the record number), and may has the following sequence.

[120] 1) Title information relevant to the lowest patent classification term is checked for the number of dots, and the title of a patent classification term relevant to a record number equal to the reference term's record number minus 1 (an immediately higher IPC term) is checked for the number of dots. It is determined that the term in question is the reference term's parent if the number of dots of the term in question is smaller than the number of dots of the reference term, it is determined that the term in question has the same level as the reference term if the numbers of dots are the same, and it is determined that the term in question is not an immediately higher or lower term if the first number is larger than the second number. As a result, only the reference term's parent is stored. If a patent classification term above the subgroup is met, the patent classification term above the subgroup is stored as the reference term's parent.

[121] 2) If the number of dots in the title of a patent classification term relevant to a record number equal to a reference term's record number minus 1 is identical to the number of dots of the reference term, the first number is larger than the second

number, or a patent classification term in question is relevant to or above the subgroup, a parent relationship is ignored, and the item 2) is performed for the patent classification term relevant to the record number equal to the reference term's record number minus 1.

[122] Although, in the above two embodiments, the methods of creating a hierarchical patent classification term DB 130 have been presented, these are only embodiments, and it will be apparent to those who develop computer systems that a reference term's parent information or a reference term's child information for the entire IPC is created from the IPC patent classification term and its relevant title information.

[123] In the case in which there is a given patent classification term, the hierarchical patent classification term DB creation module 142 creates a hierarchical patent classification term DB 130. An embodiment thereof will be described below.

[124] With respect to patent classification terms under H04B 7/00, embodiments of the results of the two methods will be presented.

[125] First, Table 2 shows the results of the method of associating each child patent classification term with a parent that is the classification term immediately above the child patent classification term, under which the child patent classification term falls (child - > parent).

[126] Table 2

[127] Next, as shown in Table 3, an embodiment of results of the method in which a parent has information about all children below the parent (parent -> child) is presented.

[128] Table 3

[129] A second sub-module is a lower patent classification term processing module 144, and functions to search the hierarchical patent classification term DB 130 for lower patent classification terms that fall under a patent classification term that is obtained from a search query input by a user or the like.

[130] In order to include patent documents corresponding to one or more patent classification terms below an obtained patent classification term in search results, one or more of the following three tasks is required. These tasks are performed by the lower patent classification term processing module 144.

[131] First, all lower patent classification terms of the obtained patent classification term are searched for based on the time of the obtainment using the hierarchical patent classification term DB 130. Since the patent classification terms in the hierarchical patent classification term DB 130 are organized in a tree structure, the lower patent classification terms can be searched for. In the case in which a specific patent classification term is included in an obtained search query, all patent classification terms below the above patent classification term can be searched for based on the tree structure, and the results of the research are transmitted to a modified search query creation module. The modified search query creation module creates a search query in which the received patent classification terms have been considered, and queries the search engine 110 using the search query.

[132] Second, the lower patent classification terms of each patent classification term are

searched for in advance, the found lower patent classification terms are stored (in a matching table) in association with the patent classification term, and lower patent classification terms associated with a specific patent classification term can be immediately transmitted to the modified search query creation module in the case where an obtained search query includes the specific patent classification term.

[133] The ranges of lower patent classification terms extracted by the lower patent classification term extraction module are classified into two types. One type of range is based on a method of extracting only patent classification terms immediately below an obtained patent classification term, while the other type of range is based on a method of extracting all lower patent classification terms that fall under an obtained patent classification term. Even in the case in which only immediately lower patent classification terms are extracted, all lower patent classification terms that fall under an obtained patent classification term may be extracted in such a way that immediately lower patent classification terms are, in turn, extracted for each of the extracted immediately lower patent classification terms, and the above process is repeated.

[134] For example, when a patent classification term included in a search query is H04B

7/15, immediately lower patent classification terms are H04B 7/155, H04B 7/185 and H04B 7/204, and all lower patent classification terms are H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7/195, H04B 7/204, H04B 7/208, H04B 7/212 and H04B 7/216.

[135] It will be apparent to those skilled in the art that the above-described extraction of immediately lower patent classification terms and all lower patent classification terms can be directly and recursively performed in the child -> parent association relationship or the parent -> child association relationship.

[136] The lower patent classification term extraction module combines extracted lower patent classification terms together and then stores them. When a patent classification term included in a search query is H04B 7/15, "H04B 7/15, H04B 7/155, H04B 7/185, and H04B 7/204" are combined together and are then stored in the case in which only immediately lower patent classification terms are stored, and "H04B 7/15, H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7/195, H04B 7/204, H04B 7/208, H04B 7/212 and H04B 7/216" are combined together and are then stored in the case in which all lower patent classification terms are stored together.

[137] Third, a modified search query is created by integrating the combined patent classification terms with the original existing search query using the modified search query creation module. The integration uses different modification methods depending on whether a search is a range search or not. First, the case in which a search is not a range search will be described first.

[138] At the location of the patent classification term in the existing search query

including the patent classification term, the patent classification term and the extracted and combined lower patent classification terms may be bound using the 'OR' relationship, the patent classification term of the original existing search query may be replaced with the bound patent classification terms. In this case, it is apparent that, when two or more patent classification terms are included in the original existing search query, the integration of patent classification terms can be performed for each of the patent classification terms.

[139] For example, when a given search query is "{key word=wireless and active} and

{applicant=Samsung Electronics} and {IPC=H04B 7/15}," the modified search query creation module searches for IPC H04B 7/15 in the given search query, and creates the modified search query "{key word=wireless and active} and {applicant = Samsung Electronics} and {IPC = H04B 7/15 OR H04B 7/155 OR H04B 7/185 OR H04B 7/204}" or "{key word= wireless and active} and {applicant = Samsung Electronics} and {IPC = H04B 7/15 OR H04B 7/155 OR H04B 7/165 OR H04B 7/17 OR H04B 7/185 OR H04B 7/19 OR H04B 7/195 OR H04B 7/204 OR H04B 7/208 OR H04B 7/212 OR H04B 7/216}", into which classification terms are combined, by obtaining immediately lower patent classification terms or all lower patent classification terms and combining them together. The latter search query is more preferable.

[140] In the summary of the functions of the lower patent classification term processing module 144, when a patent classification term is include in a search query, the lower patent classification term processing module 144 searches for information related to the lower patent classification terms of the patent classification terms, and directs the modified search query creation module to perform processing. Of course, when the patent classification term included in an obtained search query is an IPC term, the lower patent classification term processing module 144 does not need to search for a lower patent classification term in the case in which processing can be performed using a truncator, like the case of the main group or its higher class (for example, when only H04B 7/00 or H04B is obtained, it is possible to process it in the form of H04B7/* or H04B*). Accordingly, the lower patent classification term processing module 144 is more useful in the case in which there are one or more dots in title information, that is, the case of the subgroup or its lower class, for the IPC. Furthermore, it can be effectively used in other classification term systems in which the numbers of dots are used to determine positions in the classification systems.

[141] Fig. 2 sequentially shows the above-described function, which is performed by the patent search service provision system 100 of the present invention. The patent search service provision system 100 receives a search query in which a patent classification term is included from a user's computer at step Sl 10. In this case, when the title information of a patent classification term included in the obtained search query includes

one or more dots, lower patent classification terms are obtained through the patent classification term processing engine 140 in a patent classification term system at step S 120, the obtained patent classification terms are bound using 'OR' at step S 130, a modified search query including the lower patent classification terms bound using 'OR' is created at step S 140, and a search is performed using the modified search query at step S 150. The respective steps are the same as described above.

[142] When there is insertion, deletion or relocation in a patent classification term system, insertion may be processed in the same manner as in the case in which a new word is added at a specific location in a dictionary, deletion may be processed in a manner so as to eliminate the word, and relocation may be processed in a manner so as to process the case in which insertion and deletion are combined together. Of course, whenever there is insertion, deletion or relocation, the entire DB may be updated or reconstructed for the entire changed patent classification term system.

[143] The above description has been given with emphasis on the case in which lower patent classification terms are bound using 'OR' into a modified search query. Meanwhile, the typical search engine 110 supports range searches. Next, a method and system for processing patent classification terms in order to support range searches, other than the method of searching lower patent classification terms and binding them using 'OR' is presented. The present method does not search for respective lower patent classification terms of an obtained patent classification term one by one, but searches a range in which the lower patent classification terms are arranged in the patent classification term system. In particular, the range search guarantees a faster search speed for searches for data arranged in dictionary order.

[144] The reason why the range of lower patent classification terms in the patent classification term system is searched for is that typical search engines support range searches, and range searches are faster than searches for respective terms that fall within the range. For example, in the case of H04B 7/15, a considerably fast response for search results can be received when H04B 7/15, and H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7/195, H04B 7/204, H04B 7/208, H04B 7/212 and H04B 7/216 are bound using an 'OR' relationship and then the search engine 110 is queried, compared to when the search engine 110 is queried using the query 'From H04B 7/15 to H04B 7/216'. In particular, the difference in the search speed increases in proportion to the number of lower patent classification terms.

[145] However, the representative of preconditions for performing a range search is that patent classification terms existing in the search index 120 searched by the search engine 110 must be arranged in the dictionary order. As a result, if patent classification terms are arranged in the dictionary order, the classification terms may be input to the search index 120 without modification. Otherwise, a separate classification system, in

which classification terms are arranged in the dictionary order and which is input to the search index 120, must be constructed. The former is called a direct input type, while the latter is called a modified input type. A representative direct type-patent classification term system is the IPC, while a representative modified input type-patent classification term system is the USPC. Of course, it is apparent that the IPC may be converted into a modified input type through slight modification. The separate classification system is referred to as a modified patent classification term system.

[146] Fig. 3 shows another embodiment of the patent search service provision system 100 according to the present invention, which is characterized in that the terms of the patent classification term system are not used for at least one patent classification term system selected from among the IPC, the USPC, the FI, the FT and the ECLA, but modified patent classification terms are used so as to support range searches.

[147] The modified patent classification term processing engine 160 of the system of the present embodiment functions to create a modified patent classification term system, process lower patent classification terms using the modified patent classification term system, and create a modified search query by modifying an obtained search query. The modified patent classification term processing engine 160 includes a hierarchical modified patent classification term DB creation module 162, a modified patent classification term utilization lower patent classification term processing module 164, and a modified search query creation module.

[148] In order to more efficiently support range searches, a direct input type system may be modified into a modified type system using a matching table. For example, the IPC classification terms H04B 7/005, H04B 7/01 and H04B 7/185 may be respectively modified into H04B70000050, H04B70000100 and H04B70000185. Although the rules of modification may be various, the present embodiment employs a method of assigning four digits to the group and four digits to the subgroup (of course, it is preferable to accept part of the classification term system related to the section and the class, which is part of the classification term system above that related to the group). Of course, in the case in which there are extended terms below the subgroup, as in the FI, modified input type-patent classification terms may be created using a method of assigning a relevant number of digits thereto. Accordingly, when there is a direct input type-patent classification term system, the matching table can create a modified patent classification term system using only simple rules.

[149] Meanwhile, in the case in which a system has no direct input type-patent classification term system, as in the USPC, a matching table may be created. A method of creating a matching table for the USPC will be described later. The creation of the above-described modified patent classification term system is performed by the hierarchical modified patent classification term DB creation module 162. The modified

patent classification term system created by the hierarchical modified patent classification term DB creation module 162 is stored in the hierarchical modified patent classification term DB 135.

[150] The modified patent classification term processing engine 160 assigns each classification term of a modified patent classification term system to a classification term of a given classification system in a 1 : 1 association, and helps the indexer construct the search index 120 based on the modified patent classification term system. When a search query is input based on the user's given patent classification term system, the modified patent classification term utilization lower patent classification term processing module 164 searches for a modified patent classification term corresponding to the input patent classification term, the modified patent classification term utilization modified search query creation module modifies a search query using the modified patent classification term, and search results are obtained by performing a search using the modified search query. It is not necessary for the modified patent classification term to be represented in the search results.

[151] The process of performing a range search will be described in greater detail below.

It is apparent that the present range search can be applied to other patent classification term systems, such as the USPC, as well as the IPC. In order to automatically search for lower patent classification terms that fall under a given patent classification term and performs a range search, it is important to search for the next sibling of the given patent classification term in the tree structure thereof. This is the same as the modified patent classification term.

[152] When a representation of each subgroup is handled as a character string other than numerals, the IPC classification system may be considered to be a representative direct input type-patent classification system. As in a dictionary, H04B 7/01 comes after H04B 7/005, H04B 7/12 comes after H04B 7/02, and H04B 7/19 comes after H04B 7/185. In this case, when there is H04B 7/02 in an obtained search query, H04B 7/04, H04B 7/06, H04B 7/08, H04B 7/10 and H04B 7/12 below H04B 7/02 are searched for in the hierarchical patent classification term DB 130 having a tree structure. A method of recursively performing a search immediately before the next sibling may be employed as a method for the searching. An example of the method will be described in detail below. When there is a specific patent classification term, a tree structure is searched for the term's parent having children and the children are searched for the term's immediately next child. In greater detail, the next sibling of H04B 7/02 is H04B 7/15. Here, in order to find the lower patent classification terms of H04B 7/02, H04B 7/15, which is a next sibling, may be first searched for, and then all patent classification terms in a range greater than H04B 7/02 and less than H04B 7/15 are searched for next.

[153] In this case, when the next sibling exists at a node/layer other than the term's immediately higher parent node/layer (for example, the next sibling of H04B 7/15 (having two dots) does not exist immediately below H04B 7/14, which corresponds to the immediately higher parent layer/node, but is H04B 7/22 (having one dot), which exists below an immediately, immediately higher parent), the next sibling is a null value, in which case the next sibling of a parent layer/node is recursively searched for until the parent node/layer is not null.

[154] Next, the USPC will be described in greater detail below. In the USPC, which is the patent technology classification system of the USPTO, it is difficult to know the hierarchical relationship or correlation between respective subclasses using only the numbers of the subclasses without requiring a dot structure. That is, since the USPC is not organized according to the dictionary order, unlike the IPC, more systematic modification is required. The USPC or the USPC patent classification system (http://www.uspto.gov/web/patents/classification/... or classification Index to U.S. Patent Classification (a.k.a., Classification Index File) File), which was disclosed in U.S. www.uspto.gov in June of 2006, may be considered to be a modified input type. In this case, USPC codes, which can be visually recognized, are not arranged in the dictionary order, unlike the IPC codes, and thus modification is required.

[155] The following two methods may be taken as examples of the method of the modification.

[156] First, a tree structure is constructed for the entire USPC on the basis of title information including dot structures. The tree structure must be configured such that the USPC can create the nodes of the tree structure in which the parent-child relationships between USPC codes have been considered. While the USPC tree structure is searched using a depth-first search technique, modified patent classification terms arranged in the dictionary order are attached to respective nodes of the tree structure. Of course, it is natural to store the USPC codes of respective nodes in a modified patent classification DB in association with respective modified patent classification terms relevant to the nodes.

[157] Second, after a tree structure has been created for the entire USPC system based on the Index to U.S. Patent Classification File (this file includes a USPC code, a description, information about the number of dots (depth information based on a root) and information about a reference term's parent for each title) provided by the USPTO, modified patent classification terms are attached to respective nodes in the dictionary order using a search technique such as a depth-first search technique, and modified patent classification terms for the respective titles are stored in association with relevant USPC codes.

[158] Since the two examples are only embodiments, methods of associating respective

USPC codes with modified patent classification terms using title information including the dot structure of the USPC and organizing the system of the modified patent classification terms in the dictionary order may vary according to the modified algorithm. An example of such methods is a method of organizing the IPC into a table in which title information has been considered and creating modified patent classification terms, arranged in the dictionary order, through the table. Accordingly, it is impossible to describe all of the above modification methods, and furthermore, such modifications will be apparent to those skilled in the art. It will be apparent that modified patent classification term systems modified using the apparent algorithms are included in the scope of the present invention.

[159] A detailed description will be given with reference to actual examples below.

[160] The following description is related to a first part of the subclass of US class 002 apparel.

[161] The description shows a part that is copied from http://www.uspto.gov/web/patents/classification/uspc002/sche d002.htm. Although the class number 002, preceding the following numbers, is omitted, the class number 002 must be considered to be attached to each of the following numbers.

[162] 1 MISCELLANEOUS

[163] 455 GUARD OR PROTECTOR

[164] 456 • Body cover

[165] 457 ## Hazardous material body cover

[166] 458 •• Thermal body cover

[ 167] 2.11 ## Body cover for astronauts

[168] 2.12 ### Having relatively rotatable coaxial coupling component

[169] 2.13 ### Having corrugated component

[170] 2.14 ## Body cover for aviators

[171] 2.15 •• Underwater body cover for divers

[172] 2.16 ### Having an insulation layer

[173] 2.17 ### Having a garment closure

[174] 459 • Shoulder protector

[175] From the dot structure, it can be seen that 457, 458, 2.11, 2.14 and 2.15 fall immediately below 456 Body cover, and that 2.12 and 2.13 fall immediately below 2.11. It will be apparent that the above-described part can be organized into a table which takes dot structures into account. Since the relevant description thereof has been provided in the description of the IPC, it is omitted in the present description of the USPC.

[176] A tree structure in which the dot structure has been considered and into which the

USPC is organized is shown in Fig. 5. Fig. 6 shows the fact that modified patent clas-

sification terms that can be arranged in the dictionary order can be associated with respective nodes of the USPC tree shown in Fig. 5. Fig. 6 shows the fact that the symbol "=>" and relevant modified patent classification terms can be assigned to respective USPC codes in the tree structure. In Fig. 6, an indication of information corresponding to the class indication "002" is omitted in each modified patent classification term. The indication corresponding to the class indication "002" may be provided in various forms, such as "002/", "0020," "USPC+002" and "002." Fig. 7 shows the fact that the modified patent classification terms can be associated with respective USPC coded in a 1:1 correspondence. Fig. 7 shows the fact that "0020" is selected from among "002/", "0020," "USPC+002" and "002" as an indication corresponding to the class indication "002." That is, through the processing illustrated in Figs. 6 and 7, respective USPC codes can be associated with modified patent classification terms arranged in the dictionary order in a 1:1 correspondence, and the modified patent classification terms also have the same tree structure as the USPC due to the characteristics of the 1 : 1 correspondence. Fig. 8 shows the fact that a tree structure, such as that shown in Fig. 5, can be created using modified patent classification terms. That is, it can be seen that the tree structure of Fig. 5 and the tree structure of Fig. 8 have completely corresponding structures. Through the above process, respective USPC codes can be stored in association with modified patent classification terms. An example of the storage may be the following table. The following table 4 may be an embodiment of the modified patent classification DB according to the present invention.

[177] Table 4

[178] The Index to U.S. Patent Classification provided by the USPTO will be described.

The first three digits of nine digits in a first column designate a class number, and the remaining six digits designate a subclass number. Although there is a decimal point or a point, as in 2.11, in an Internet site, it can be seen that it is represented in the form of 002110. The subsequent digit designates a depth relative to a root. 'APPAREL,' representing the class 002, has a depth of 0, 'MISCELLANEOUS,' having no dots, has a depth of 1, which indicates that 'MISCELLANEOUS' is located immediately below 'APPAREL', and 'Thermal body cover,' having two dots, has a depth of 3. The column next to the depth column shows a serial number, and the next column thereafter shows a number indicating a parent. That is, in "455000 body cover," 455000 attached beside 'body cover' indicates that its parent has 455000, and that 'GUARD OR PROTECTOR' is its parent. This can be considered to be a representation corresponding to the dot structure. As a further example, in "456000 Body cover for aviators," it can be seen that 'Body cover for aviator's parent is 'Body cover,' which uses 456000.

[179] Since the above-described USPC classification terms are not arranged in the dictionary order, it is impossible to perform a range search unless re-processing is performed.

[180] It will be easy for those skilled in the art to creates a tree structure, such as that shown in Fig. 5, using parent information and relative depth information in each USPC code contained in the Index to U.S. Patent Classification File, and it will be simple to form the tree structure of modified patent classification terms, such as that shown in Fig. 8, using the tree structure. That is, it will be apparent that it is easy to create a tree structure, such as that shown in Fig. 5, using information provided by the Index to U.S. Patent Classification and assign modified patent classification terms to the tree structure, as shown in Fig. 6, and it is possible to form a tree structure composed of modified patent classification terms, such as those shown in Figs. 7 and 8, using the modified patent classification terms. Through the above process, respective USPC codes can be stored in association with modified patent classification terms. An

example of the storage may be the following Table 5. The following table may be an embodiment of the modified patent classification DB according to the present invention. Table 5

[182] Of course, the USPC system including given title information can be associated with modified patent classification terms arranged in the dictionary order using an algorithm different from the above algorithm, and it will be apparent that the above association can be stored in the form of a matching table of modified patent classification terms that is associated with the USPC, as described above.

[183] Fig. 4 shows the flow of an embodiment in which the hierarchical modified patent classification term DB creation module 162 creates the hierarchical modified patent classification term DB 135.

[184] The hierarchical modified patent classification term DB creation module 162 obtains all patent classification terms in a patent classification term system at step S310, arranges all of the obtained patent classification terms in a tree structure with reference to the dot information of titles at step S320, associates respective patent classification term nodes with modified patent classification terms while searching a patent classification term system, organized in the tree structure, using a depth-first search technique at step S330, and stores the relationship between respective patent classification terms and relevant modified patent classification terms in the hierarchical modified patent classification term DB 135 at step S340.

[185] Next, the method in which the modified patent classification term utilization lower patent classification term processing module 164 of the present invention processes obtained patent classification terms and obtains information about modified patent classification terms to create a modified search query with focus on a range search will be described.

[186] In the above case, when the USPC "002/456" is obtained from an input search query, 002/456, 002/457, 002/458, 002/2.11, 002/2.12, 002/2.13, 002/2.14, 002/2.15, 002/2.16 and 002/2.17 must be searched for so as to perform a lower classification inclusion search. Here, as seen from the tree structure of Fig. 5, the next sibling of 002/456 is 002/459. Accordingly, when association with modified patent classification terms is performed in Table 5 and an instruction to search for all modified patent classification terms above or even with 002002100 associated with 002/456 (the 'above' means 'later' in the patent classification term/modified patent classification term system organized in the dictionary order, which can transmit a clear meaning to those skilled in the art in conjunction with a range search without requiring an additional explanation) and below 002002200 associated with the next sibling 002/459 is given, a lower classification inclusion search is performed.

[187] In greater detail, there are two methods as follows. The first method is a method of, when "002/456" is obtained, searching for the relevant modified patent classification term 002002100, searching for 002/459, that is, the next sibling of 002/456, searching for the modified patent classification term 002002200 associated with 002/459, and setting a range based on 002002100 and 002002200. The range may be greater than or equal to 002002100, and less than 002002200. The characteristic of the present method is to first search for the next sibling of an obtained patent classification term in a tree structure before modification and then to search for modified patent classification terms respectively associated with the obtained patent classification term and the next sibling. In the present method, the next sibling of a patent classification term input by a user through a search query is first searched for, and modified patent classification terms respectively associated with the patent classification term and the next sibling of the patent classification term are then searched for, based on the fact that a modified patent classification term system is associated with a relevant patent classification term system in a 1 : 1 correspondence from the point of view of the structure.

[188] Fig. 10 shows the flow of an embodiment of a method of performing a search using a modified search query based on the first method.

[189] The patent search service provision system 100 obtains a search query in which a patent classification term is included at step S710. Thereafter, in the case in which the obtained patent classification term has one or more dots in the title information thereof, the modified patent classification term utilization lower patent classification term processing module 164 (or the lower patent classification term processing module 144) searches for the next sibling patent classification term of the obtained patent classification term in a tree structure in which a hierarchical patent classification term system has been considered at step S720. Thereafter, the modified patent classification term utilization lower patent classification term processing module 164 obtains a modified patent classification term and a next sibling modified patent classification term, respectively associated with the obtained patent classification term and the next sibling patent classification term, from the hierarchical modified patent classification term DB 135 at step S730. Thereafter, the modified search query creation module creates a modified search query including a range based on the modified patent classification term and the next sibling modified patent classification term of the modified patent classification term at step S740, and the search engine 110 performs a search using the modified search query at step S750.

[190] The second method is a method of, when "002/456" is obtained, searching for the relevant modified patent classification term 002002100, searching a modified patent classification term system for the next sibling 002002200 associated with 002002100, and setting a range based on 002002100 and 002002200. The range may be greater

than or equal to 002002100 and less than 002002200. The characteristic of the present method is to search for a modified patent classification term associated with an obtained patent classification term and search for a next sibling associated with the modified patent classification term. That is, in the present method, a modified patent classification term for an input patent classification term is first searched for, and the modified patent classification term system is searched for the next sibling of the found modified patent classification term. The second method is advantageous in that the number of steps is one less than that of the first method.

[191] Fig. 9 shows the flow of a method of performing a search using a modified search query based on the second method.

[192] The patent search service provision system 100 obtains a search query in which a patent classification term is included at step S510. In the case in which the obtained patent classification term includes one or more dots in the title information thereof, the modified patent classification term utilization lower patent classification term processing module 164 obtains a modified patent classification term associated with the obtained patent classification term from the hierarchical modified patent classification term DB 135 at step S520, and searches for the next sibling modified patent classification term of the modified patent classification term in a tree structure in which the modified patent classification term system has been considered at step S530, the modified search query creation module creates a search query including a range based on the modified patent classification term and the next sibling modified patent classification term of the modified patent classification term at step S540, and the search engine 110 performs a search using the modified search query at step S550.

[193] Through the first or second method, the modified patent classification term utilization lower patent classification term processing module 164 can find two modified patent classification terms that correspond to both limits of a range that is used to perform a range search on an obtained patent classification term.

[194] It is apparent that information about the modified patent classification terms associated with respective USPC codes must be included in the search index 120. Although each of the modified patent classification terms may be formed of a number having a specific number of digits, it may be formed to include a number and one or more characters. Meanwhile, when there is an addition to the USPC, a modified patent classification term can be created so that the dot structure is considered, an overlap does not occur, and the term has an appropriate position when arranged in the dictionary order. A method of creating a position in the dictionary order corresponds to a technique identical to a well-known technique for, when a new word is added to a dictionary, appropriately positioning the new word in alphabetical order, and can be easily implemented by those skilled in the art. Furthermore, when more lower classes

are newly created when positioning is performed in the dictionary order, an extension can be continuously added as a last digit. For example, when a plurality of patent classification terms is hierarchically created below 002/2.12, modified patent classification terms can be continuously created, like 0020021310a and 0020021310b, below the modified patent classification term 002002131 associated with 002/2.12 and above 002002132 in the dictionary order, while an extension is continuously added as a last digit.

[195] The fact that a range search can be performed by introducing modified patent classification terms, even in the case of a patent classification term system that does not have its own dictionary order arrangement structure, like the USPC, has been described. Of course, it will be apparent that patent classification terms that already have a dictionary-order arrangement structure, like those of the IPC, can also be associated with new modified patent classification terms. Accordingly, the spirit of the present invention can be applied to the FI and the ECLA, which are extensions of the IPC, and FT patent classification terms. That is, in order to perform range searches in the IPC, the FI, the FT, the ECLA, the USPC or the like, modified patent classification terms in which the dot structures of respective classification terms are considered can be created, and the modified patent classification term can support range searches because they are arranged in the dictionary order.

[196] It is preferred that there be information into which original patent classification terms have been processed and information into which modified patent classification terms have been processed together in the search index 120 for the search engine 110. If a user does not want search results for lower patent classification terms, it is not necessary to use an index in which modified patent classification terms have been processed into information.

[197] Thereafter, the modified search query creation module of the present invention creates a modified search query by incorporating the obtained modified patent classification terms with an original existing search query.

[198] With regard to a range search, the components of a modified search query vary slightly depending on whether the search engine 110 supports heterogeneous range term processing for the range. The term 'heterogeneous range term processing' refers to the processing of the case in which the types of both terms that define a range are different, like "a range greater than or equal to 'A' and less than B." In the case in which heterogeneous range term processing is supported, the modified search query component "(patent classification terms in a range greater than or equal to the obtained patent classification term and less than the next sibling of the obtained patent classification term)" is created for an obtained patent classification term. If heterogeneous range term processing is not supported, the search query component "{(the obtained

classification term) or (patent classification terms in a range greater than or equal to the obtained patent classification term and less than the next sibling of the obtained patent classification term)}" is created. Of course, it will be apparent that, when modified patent classification terms are arranged in the dictionary order, the modified patent classification terms are also processed as described above.

[199] If an obtained search query is "{key word=wireless and active} and {applicant =

Samsung Electronics} and {IPC = H04B 7/15}" and heterogeneous range term processing is supported, "{key word=wireless and active} and {applicant = Samsung Electronics} and {IPC >= H04B 7/15 and IPC <H04B 7/22}" is obtained. In contrast, if heterogeneous range term processing is not supported, "{key word=wireless and active} and {applicant = Samsung Electronics} and {IPC = H04B 7/15 or IPC > H04B 7/15 and IPC <H04B 7/22}" is obtained.

[200] The gist of the above-described patent classification term processing engine 140 of the present invention resides in the function in which, when a given patent classification term exists, automatically including all children or patent classification terms immediately below the patent classification term and producing search results in which the lower terms have been considered.

[201] The construction of the patent search service provision system 100 of the present invention using the above-described patent classification term processing engine 140 and a method of searching patent information using the system and processing patent information based on the found patent information will be described below.

[202] First, the method of searching for patent information using the patent search service provision system 100 of the present invention is presented. The gist of the patent information search method of the present invention resides in the fact that, when there is a patent classification term in a search query, patent information including patent classification terms below the patent classification term are also provided.

[203] A method of processing information when a patent classification term below the main group is included in a search query is presented as a first embodiment of the search for patent information.

[204] If there is a patent classification term below the main group (At least one dot is included in title information for a patent classification term below the main group. Accordingly, it will be apparent that the spirit of the present embodiment can be applied to the USPC, the FI, the FT, the ECLA or the like, having dot structures.), the patent classification term is obtained by the patent classification term processing engine 140, the hierarchical patent classification term DB 130 is searched, information about all patent classification terms below the patent classification term given in the search query is searched for, a modified search query is created by incorporating the found terms with the obtained patent classification term, and a patent search is performed

using the modified search query. Such a patent search using a search query including a modified patent classification term is referred to as a "modified patent classification term inclusion search" in the present specification.

[205] Of course, if a patent classification term included in a search query is a main group term, 700' is separated from the main group term, and all patent classification terms below a relevant main group are searched for using a truncator. For example, if IPC = H04B 7/00 is given, IPC = H04B 7/00 is converted into IPC = H04B 7/? and all patent classification terms starting with H04B 7 are searched for. However, of course, this is not the characteristic of the modified search query creation method of the present invention, and is not considered to be an important factor. The spirit of the present invention is significantly used for patent classification term levels (patent classification term levels generally having dot structures) at which the hierarchical relationships between patent classification terms cannot be found using patent classification terms themselves.

[206] In the case in which search results are obtained using a modified search query, a display of search results may be clustered and then provided according to a searcher's selection. 'Clustering' refers to a method of putting search results together in groups. Clustering is applied to the present invention as follows. If there is a clustering window showing a clustered structure, or a clustering level can be selected, only the search results for the dot subgroup, corresponding to a clustering level set or selected in advance, and its lower subgroups can be represented. When a searcher selects a lower or higher clustering level (the clustering level is associated with a dot level), search results corresponding to the selected clustering level are represented.

[207] Taking search results related to H04B 7/00 as an example, when 1-dot subgroup level clustering is selected, 7/005, H04B 7/01, H04B 7/015, H04B 7/02, H04B 7/14, H04B 7/22 and H04B 7/24, relevant to the 1-dot subgroup, are shown in the clustering window in the form of a directory folder below H04B 7/00, and all search results relevant to H04B 7/00 are shown. In this case, when a searcher selects H04B 7/14, H04B 7/145 and H04B 7/15, which is immediately below H04B 7/14, are shown in the form of a lower directory folder, in which case only search results relevant to H04B 7/14 and its lower patent classification terms are shown. When the searcher selects H04B 7/15, H04B 7/155, H04B 7/185 and H04B 7/204, which are below H04B 7/15, are shown, and only search results relevant to H04B 7/15 and its lower patent classification terms are shown.

[208] In this case, for a searcher's convenience, the number of patent classification terms immediately below a reference term and/or the number of all lower patent classification terms are shown beside each clustering item of a clustering window (the folder name of a directory). The numbers of patent documents relevant to the patent

classification term and all of the lower patent classification terms of the patent classification term may be shown. In this case, the numbers of both applied documents and registered documents may be shown, or the number of applied documents or registered documents may be shown. For example, when the patent classification term is H04B 7/15, immediately lower patent classification terms are H04B 7/155, H04B 7/185, and H04B 7/204, so that "3" is shown. Meanwhile, since all lower patent classification terms are H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7/195, H04B 7/204, H04B 7/208, H04B 7/212 and H04B 7/216, "10" is shown. It is preferable to show both numbers. The reason for this is that there is a small number of classes immediately below a reference term, while there is a large number of classes below the immediately lower terms. This phenomenon is noticeable particularly in emerging technology fields or advanced technology fields, in which the speed of subdivision of technology or the speed of increase in the detail of classification of technology is high. If the number of immediately lower classes is small when only the number is shown, there may be users who ignore the lower classes because they cannot be aware of the importance of relevant technology fields, even though the relevant technology fields are important.

[209] Patent classification term information is organized in directory form, and the lower patent classification terms of a specific patent classification term can be systematically accessed when browsed by a user. Currently, www.delphion.com is providing such a service. However, although the website provides information about the number of immediately lower patent classification terms, 1) the website does not provide the total number of lower patent classification terms, 2) does not provide the number of patent documents for all patent classification terms, and 3) does not provide patent documents including all immediately lower patent classification terms relevant to the provided number when the number is clicked on. The present invention can solve these problems using the hierarchical patent classification term DB 130 and the modified patent classification term search method.

[210] When the patent classification term information is organized in directory form, a symbol, such a color or an icon, is indicated in each directory, so that users' ease of access can be improved. That is, in the cases in which the number of immediately lower classes is large, the number of all lower classes is large, depth is high, or the number of relevant patent documents is large, different colors or preset icons are assigned, with the result that users can know the internal characteristics of a relevant directory (the number of all lower classes, the depth of classification, and the number of patent documents) in detail. In particular, the case in which the number of all lower classes of a specific patent classification term directory is large or the depth of classification is high provides some evidence of the fact that many patent applications have

been submitted in the relevant field and patent applications have been submitted for various technologies, and thus the directory in question may be more important than other directories.

[211] In the above description, the method and system for automatically including and searching for lower patent classification terms using the search engine 110 has been described. However, it is apparent that the spirit of the present invention can also be applied to the case in which a search is performed using a DBMS as well as the case in which the search engine 110 is used. A search engine uses search queries for a search engine, while a DBMS uses DB queries for a DBMS with respect to a DB. The case in which a DBMS is used will be described in detail below. However, descriptions identical to those in the case in which the search engine is used will be omitted here.

[212] The modification of a DB query for a DBMS is not different from the modification of a search query from the point of view of the application of the spirit of the present invention. That is, when a DBMS is used, the DBMS 170 is replaced with the search engine 110 and the search index 120 in Figs. 1 and 3. The relevant construction is shown in Figs. 11 and 12.

[213] Although the spirit of the present invention has been described in conjunction with the case in which the search engine is queried using a modified search query automatically including lower patent classification terms, this is not substantially different from the case in which the DBMS 170 is queried using a modified query (which may be referred to as a modified search query for a DBMS) automatically including lower patent classification terms. That is, the spirit of the present invention can be applied to the DBMS as well as the search engine in the same manner.

[214] The method of creating modified search queries using the lower patent classification term processing module 144, the hierarchical patent classification term DB, the modified search query creation module 146, the modified patent classification term utilization lower patent classification term processing module 164 and the hierarchical modified patent classification term DB 135 of the present invention with focus on the search engine 110 is used without modification for the creation of modified queries for a DBMS. That is, the patent classification term processing engine 140 and modified patent classification term processing engine 160 of the present invention are not dependent on the search engine 110 or the DBMS 170, but may be considered to perform pre-processing for the information processing of the search engine 160 or the DBMS 170. That is, the patent classification term processing engine 140 and modified patent classification term processing engine 160 of the present invention perform preprocessing independent of the search engine 110 or the DBMS 170, and the results of the pre-processing are considered in the hierarchical patent classification term DB 130 and the hierarchical modified patent classification term DB 135. The results of the pre-

processing considered in the hierarchical patent classification term DB 130 and the hierarchical modified patent classification term DB 135 are considered in the search index 120 or the patent information DB 150. Accordingly, when the search engine 110 performs a search with regard to the search index 120, the results of the pre-processing have been considered in the search index 120 and/or the patent information DB 150, so that it is enough that a search query created by a user has been modified. In the same manner, if only a search query has been modified when the user queries the patent information DB 150 through the DBMS 170, the DBMS 170 outputs search results consistent with the modified search query.

[215] It is preferred that, in the case in which the DBMS 170 is used, as shown in Figs. 11 and 12, the contents of the hierarchical patent classification term DB 130 and the contents of the hierarchical modified patent classification term DB 135 be physically or logically associated in the patent information DB 150. Logical association means that tables can be logically associated with each other through instructions supported by the DBMS, even though they are separate from each other. In this case, an instruction to join tables that can be logically associated with each other may be included in the supporting instructions. This scheme will be apparent to those skilled in the art who deal with a DBMS.

[216] The modified search query creation module 146 creates a search query (DB query) including the lower patent classification terms of a patent classification term by processing the patent classification term that is included in the search query (DB query) obtained from a user. That is, in the case in which a DBMS is queried using a search query for a DBMS, a patent classification term is included in the search query for a DBMS and one or more dots are included in the title information of the patent classification term, the lower patent classification terms of the patent classification term are automatically obtained and then a modified query for a DBMS can be automatically created. That is, 1) the above-described method of searching for a lower patent classification term and forming a range based on the patent classification term and the lower patent classification term using 'OR' may be used as it is, and 2) the above-described method related to the modified patent classification term may be used as it is. The DBMS may be queried using a modified query in the same manner as the search engine is queried using the modified search query.

[217] Accordingly, it is apparent that the processes of Figs. 2, 4, 9 and 10 are applied to the case in which the DBMS 170 is used in the same manner.

[218] The patent information DB 150 includes bibliographic information, body text and drawing information, obtained from patent documents for each country. In particular, various types of information of bibliographic information, which are targets for analysis, are arranged for respective fields. Furthermore, counted information may also

be arranged in the patent information DB. The counted information may include the number of claims, the number of drawings, the number of families, the number of inventors, the number of applicants, and the number of patent classification terms as examples. Furthermore, the counted information may further include the number of foreign patent applications related to each patent document, the number of citations, and the number of times it has been cited.

[219] It is apparent that, when one or more criteria selected from among a preset country, a period, an applicant, a technical field (including a patent classification term), a specific technical key word and an inventor are added to the DB query, it is possible to obtain more detailed search results.

[220] In this case, there may be a problem with search efficiency depending on which of the first method, using a search engine, or the second method, using a DBMS, is employed. Generally, when a DBMS is used, the DBMS supports full-text searches but has the problem of low speed compared to a search engine. Accordingly, when a key word is included in a search query, it will be more efficient to find the number of patent documents consistent with the search query using the search engine.

[221] Next, a patent analysis method based on patent classification terms using a modified patent classification inclusion search technique according to the present invention will be described. The patent analysis method is performed by the patent information analysis engine 200 of the present invention. The overall analysis of patent information performed by the patent information analysis engine of the present invention includes the following two cases.

[222] The first method is an analysis method that is performed based on the number of patent documents consistent with a search query obtained by the search engine 110 when the search engine 110 is queried using a specific query including a patent classification term.

[223] The second method is a method of querying the DBMS of the present invention, and is an analysis method based on the number of obtained patent documents.

[224] The results of the first and second analysis can be obtained in a time series manner

(including specific time/a specific period), in which case the concept of speed can be acquired by dividing the results of analysis by the time/period. Analysis including the concept of speed is referred to as speed analysis. Next, the speed analysis method will be described. An engine for performing speed analysis is referred to as a speed analysis engine 210.

[225] Speed analysis based on patent classification terms refers to the acquisition of 1) numerical information and 2) personal information based on patent classification terms from patent classification term systems supported by country-based patent information for one specific country or two or more countries.

[226] The numerical information may include 1) the number, 2) increase rate in the number, 3) increase speed, and 4) increase acceleration per preset period. If the numerical information per preset period is known, the increase rate in the number can be calculated using {the number in a relevant period ? the number in the previous period}/{the number in the previous period}*100, the speed can be calculated using {the number in a relevant period ? the number in the previous period }/{ the conversion length of the period}, and the acceleration can be calculated using the variation rate in the speed. Accordingly, although the speed analysis will be described with focus on the number in the present specification for ease of description, it must be assumed that the description includes the items 2), 3) and 4). The numerical information may be the number of information items that can be derived from patent information, and may include 1) simple tallied numbers, such as the number of patent applications and the number of patent registrations, 2) rates, such as registration rate, occupation rate, and concentration rate, 3) indexes such as a patent activity index, 4) the number of persons, such as the number of applicants, the number of inventors, and the number of agents, 5) the number of countries related to a specific patent document and the number of administrative organizations from which a specific patent document can be obtained in each country (a patent office, a primary decision organization such as a patent tribunal, a secondary decision organization such as a patent court, and a supreme court), 6) various types of numbers included in each patent document, such as the number of claims, the number of pages of each specification, the number of drawings, the number of citation documents, the number of cited documents and the number of documents in an examination, 7) the number of events occurring during the lifecycle of each patent application, such as the number of specific administrative processing events, and the resulting values of each function, such as a significance evaluation function using combinations of the above items or respective items as input variables, as examples.

[227] Personal information refers to information about persons, such as 1) an applicant or a patentee, 2) an inventor, and 3) an agent, related to each patent document, which is the basis for calculation of 1) the number, 2) increase rate in the number, 3) increase speed, and 4) increase acceleration, which are the targets of the analysis.

[228] As a simple example of the speed analysis, the number of patent applications or the number of patent registrations per year may be calculated for the entire patent classification system or a specific class (for example, the H04B 7/00 group). Although the IPC is taken as an example here, a desired object can be achieved for other patent classification systems having a dot structure using a similar method. Furthermore, although the number of patent applications is described as an example of the analysis target for ease of description, a desired object can be achieved for other analysis targets using a similar method.

[229] First, the patent classification of a specific patent class will be described. The entire patent classification system is the highest set of specific patent classes, and correspon ds to all technical fields.

[230] First, information designating a specific class is obtained. If the obtained patent class is above the subclass, the patent class above the subclass is resolved into main groups because there are dot structures in the subgroup and its lower classes. Each of the resulting main groups is searched for 1-dot subgroups in subgroups (immediately lower class patent classification terms on the basis of the main group), and the modified patent classification term inclusion search of the present invention must be performed for the 1-dot subgroup and its lower classes (of course, although the modified patent classification term inclusion search of the present invention may be used because a range search can be performed for the main group and its higher classes, a truncated search can be performed for this area, but only the method of the present invention can be used for the 1-dot subgroup and its lower classes).

[231] For example, if the obtained patent class is H04B at the subclass level, H04B is primarily resolved into lower classes relevant to H04B, that is, the main groups H04B 1/00, H04B 3/00, H04B 5/00, H04B 7/00, H04B 10/00, H04B 11/00, H04B 13/00, H04B 14/00, H04B 15/00 and H04B 17/00. Furthermore, a 1-dot subgroup, which pertains to each of the main groups, in turn, is searched for the main group. Only when the modified patent classification term inclusion search method unique for the present invention is used for 1-dot subgroups and their low classes can a search including the lower patent classification terms below the 1-dot subgroup be performed.

[232] When the main class is H04B 7/00, the modified patent classification term inclusion search of the present invention is performed for all 2-dot subgroups, 3-dot subgroups and/or 4 dot subgroups with respect to each of seven 1-dot subgroups, which are levels immediately below the main class, among a total of 25 lower patent classification terms, which pertain to the main class.

[233] The number of patent classification terms that are the targets of the above analysis may be two or more. The groups may be levels (for example, all of the groups are subclasses), and may be a mixture of two different levels (some groups are subclasses, some other groups are main groups, and some other groups are classes).

[234] When the above patent classification term groups are extended, the patent information analysis engine 200 can perform processing for the entire patent classification term system. That is, the modified patent classification term inclusion search of the present invention may be performed for each of all patent classification terms corresponding to the subgroup or below the subgroup.

[235] The case in which all subgroup-level IPC terms are obtained for H04B 7/00 and the number of patent applications per year is calculated for each of the terms is taken as an

example below.

[236] In this case, an obtained search query does not include only patent classification terms, but may further include a key word or applicant information. Accordingly, a search query can produce search results thereof through an "AND" operation between the key word or applicant information and the patent classification term. The search engine 110 may transmit a search query, obtained by performing a combination for all fields associated in an index by the search engine 110, as a query. If there is a patent classification term in the search query obtained through the combination, a modified search query, obtained by applying the modified patent classification term inclusion search method to the patent classification term, can be transmitted as a query. The search engine 110 produces search results based on the modified search query. Generally, since the search engine 110 rapidly outputs the number of search results relevant to a search query (in contrast, the provision of a list of search results imposes a relatively high load on a server and a network), a desired object can be achieved without taking an excessively long time when the numbers of items are determined for statistics or analysis.

[237] The case in which, when Samsung Electronics is input as an applicant, all subgroup-level IPC terms are obtained for H04B 7/00 and the number of patent applications per year is calculated for each of the IPC terms is taken as an example.

[238] It is apparent that numerical information other than the number of patent applications per year can be calculated using the patent information analysis engine 200 of the present invention. The patent information analysis engine 200 can primarily search for numerical information easily. The numerical information includes various types of numerical information that can be searched for in the patent documents of a relevant country through direct or indirect calculation, in addition to basic numerical information, such as the number of patent applications, the number of patent registrations and the registration rate per year or preset period. The various types of numerical information include 1) information that can be provided by the search engine 110 using a combination of search queries directly or through simple processing without requiring special information processing to be conducted in advance, 2) information that has undergone special previous processing, and 3) information that can be provided through another patent information analysis engine 200.

[239] Examples of the item 1) are most bibliographic information, and representatives thereof are the number of patent applications and the number of patent registrations for each country. The registration rate can be calculated by dividing the number of patent registrations by the number of patent applications. However, the case in which the calculation of the registration rate is not performed using the number of patent registrations/the number of patent applications per year, but whether each patent ap-

plication has been registered is determined, is relevant to the item 3).

[240] A representative example of the item 3) may be ranking information for each applicant that meets a specific search condition. In order to output ranking information for each applicant, a method of obtaining a search result list, checking an applicant list included in the result list, calculating the frequency of appearances of a specific applicant, and assigning a rank proportional to the frequency may be used. In this case, the calculation of the rank is performed using a rank calculation engine.

[241] A representative example of the item 2) may be the number of significant patents.

Although the criteria for determining the significance of patents may vary, significance may be determined using, for example, a method of assuming that (1) a patent having many related foreign patent applications, (2) a patent having many claims, (3) a patent having many pages in the specification, and (4) a patent that has been cited many time have high significance, assigning weights to the respective patents, determining which items are relevant to a patent document in question and calculating the sum of the weights. When such sums of weights have already been calculated for respective patent documents and respective criteria, the number of patent documents relevant to each criterion in all search results can be determined.

[242] The search engine 110 performs a search using a received search query in the above-described manner, and returns search results including the number of search results.

[243] A method of searching patent information for significant information with a focus on the number of search results using the speed analysis method will be disclosed below. 'Speed analysis' refers to the analysis of the transition of a search target for each of the patent classification terms, which constitute a group including a patent classification term, per preset period. The 'search target' refers to an index that is the target of a search, such as the number of patent applications and the number of patent registrations, and examples of the transition are the number, the change in the number (speed), and the speed of change in the number (acceleration).

[244] In the case in which the search target is the transition of patent applications for each technical field (each patent classification term), the interpretation of the transition may vary, as described in the following examples.

[245] I) A specific technical field, having a small number of patent applications, a rapid increase in the number of patent applications and a positive (+) acceleration of increase in the number of patent applications, has a strong probability of being a new technical field that is opening up. It can be considered that there is a great need to monitor applicants and technologies related to patent applications in the field. In particular, in the case in which the same pattern exists abroad or the same pattern exists in several specific countries, there is a strong probability of a relevant field being a field in which

competition for innovation is intense.

[246] 2) In the case in which a specific technical field has a small number of patent applications, a relatively rapid increase in the number of patent applications and a negative (-) acceleration, there is a strong probability that the technical field is a niche field that has been occupied by a small number of persons, which is controlled by a small number of persons.

[247] 3) In the case in which a specific technical field has a small number of patent applications and a negative increase in the number of patent applications, there is strong probability that the field is a field in which there is little market interest, the related market is small, and the development of important technologies has been already terminated, and thus defensive patent applications are being submitted to defend the technical field.

[248] 4) In the case in which the specific technical field has a large number of patent applications, a rapid increase in the number of patent applications and a positive acceleration of increase in the number of patent applications, it can be considered that the technical field is a field for which the scale of the related market is large and that large enterprises are active because the expectation for profits in the future is high.

[249] 5) In the case in which a specific technical field has a large number of patent applications, a relatively rapid increase in the number of patent applications and a negative (-) acceleration, this may be a pattern in which relevant market participants are focusing patent applications on applied technologies rather than on claims for rights to core technologies.

[250] 6) In the case in which a specific technical field has a large number of patent applications, a low or negative (-)increase in the number of patent applications and a negative (-) acceleration in the number of patent applications, there is a strong probability that the technical field is a field for which defensive patent applications for claims for rights to processing technologies or improved technologies are being made.

[251] Accordingly, various types of useful information can be obtained by measuring the number, the change in the number (speed) and the speed of change (acceleration) for all patent classification terms relevant to the entire technical field. Through this process, a distribution curve or a distribution map based on the speed for a specific technical field or the entire technical field can be constructed.

[252] In the same manner, analysis based on patterns, like the above items 1)~6), can be performed for the number of applicants or the number of higher applicants through secondary processing, instead of the number of patent applications.

[253] In the case in which the number of applicants is the same as in item 1), that is, if the number of applicants is small, the speed of increase in the number of applicants is high and the acceleration of increase in the number of applicants is positive (+), a related

field is a technical field in which some technical enterprises are making leading investments, the interest of participants is gradually increasing due to a promising future outlook, and there is a strong probability that the return on investment will be great and profit will be high.

[254] In the case in which the number of applicants is the same as in the item 4), there is a strong probability that the related field is a field that has the characteristic of a red ocean, in which competition has become intense.

[255] In the same manner, when the above-described speed analysis is performed for the entire technical field or a specific technical field (specific patent classification term) based on a specific applicant, the following results can be expected.

[256] In the case in which a specific applicant has a small number of patent applications, a high speed of increase in the number of patent applications and a positive (+) acceleration of increase in a specific technical field, there is a strong probability that the applicant considers the field to be a new future business field and performs concentrated technical development in the field.

[257] In the case in which a specific applicant has a high standard deviation in the number of patent applications in a specific technical field, this means that technical development is difficult in the field. The reason for this is that, in the case in which technical development is advanced toward a target, the number of patent applications exhibits a gradually increasing pattern and the standard deviation in the number of patent applications tends to be low.

[258] In the case in which a specific applicant has a large number of patent applications, a high speed of increase in the number of patent applications and a low acceleration in the number of patent applications in a specific technical field, there is a strong probability that the applicant considers the technical field to have passed through a period of rapid development.

[259] With regard to the above individual index, such as the number of patent applications, the number of patent registrations or the number of applicants, or a mixture of two or more indices, the individual index or the mixture of the indices may be considered in a search query (for example, (applicant = Samsung Electronics) and (IPC = H04B 7/15)). Significant information can be obtained when the search results are arranged along the time axis or are output in the form of a table.

[260] Through the above-described speed analysis, 1) information about the tendency of patent applications for the entire technical field, 2) information about the tendency of patent application for a specific technical field, 3) information about the tendency of patent applications for the entire technical field regarding a specific applicant, and 4) information about the tendency of patent application for a specific technical field regarding a specific applicant can be analyzed. The information can be visualized

using a graph, a table or the like. Furthermore, when the item 3) and the item 4) are performed for two or more applicants and the results are compared with each other using a single table or a single graph, a comparison table, a comparison graph and a comparison map (2-dimensional or 3-dimensional; one or more of a comparison target applicant axis, a technical field axis and a year axis are included, and the height axis indicates the number) can be created.

[261] Furthermore, when there is an item that meets preset criteria, among speed analysis results for the entire technical field or a specific technical field (a technical field or a technical field for a specific applicant), it may be reported to the outside. Examples of such preset criteria may be 1) a number of patent applications equal to or greater than an average of 50 per year, a rate of increase in the number of patent applications equal to or higher than an average of 30% per year, and a rate of acceleration in the number of patent applications equal to or higher than 10%, or 2) an obtained reference value for at least one of the number of patent applications, the rate of increase in the number of patent applications and the rate of acceleration in the number of patent applications for a specific applicant and/or a specific technical field (a specific patent classification term). The reporting is performed in the form of a report, and the content of the report may include one or more of 1) material used to determine whether the criteria has been met, 2) a search query that is the target of the material used to determine whether the criteria is met, and 3) a search result patent list relevant to the meeting of the criteria for a specific applicant and/or a specific technical field that meets the criteria.

[262] The above-described speed analysis may be performed using the patent DBs of two or more countries. The cases in which 2 the patent DBs of two or more countries are involved may be basically divided into the following two series.

[263] The first series is a series in which the speed in a specific technical field is analyzed for each country. This enables technical fields, on which concentration is focused, to be found for each country. In particular, when this is associated with higher applicants having higher occupation rates for the specific technical field for each country, it can be determined which corporations lead in the technical field in the country. In the case in which there is an applicant having high occupation rates in two or more countries, there is a strong probability that the applicant is leading in the specific technology worldwide.

[264] For this purpose, the speed analysis engine 210 performs searches in such a way as to, with respect to patent DBs for respective countries, create search queries for respective countries by including the same patent classification term in search queries and query the search engine 110 using the search queries for respective countries. When there is a key word or the like, the search queries for respective countries are created by translating the key word or the like into respective country languages.

[265] The second series is a series that analyzes which applicants in a first specific country file how many patent applications in which technical fields in a second specific country. For such analysis, the object of the second series analysis can be achieved in such a way that a search query is created for a specific applicant and for the entire technical field or a specific technical field in a first country (in the case of the entire technical field, a patent classification term is not designated, and, in the case of the specific technical field, a patent classification term relevant to the technical field of the search field is designated), a search is performed, a primary list is created for the entire technical field or the specific technical field, the same search query is created for a second country (naturally, the name of an applicant is translated into the language of the second country), a search is performed, a secondary list is created for the entire technical field or the specific technical field, and the primary list and the secondary list are compared with each other for the entire technical field or the specific technical field.

[266] A method using priority right information is used as another method of achieving the object of the second series analysis. After the primary list has been created, whether a patent application number present in the primary list is present in the patent information of the second country is determined (this is possible because, when a foreign patent application is filed, a patent application is filed in a first country and then the foreign patent application is filed in a second country within a period within which a priority right can be claimed (generally, within one year from the priority date) in conjunction with the claim for the priority right). Comparison may be performed for the entire or a specific technical field with respect to the patent list having only priority information in the second country based on the results of the determination. The advantages of this method are remarkable in the case in which the representative name of an applicant is not sufficiently clear or an applicant files a patent application in the second country under the name of another party.

[267] The third series is a cross search series using different types of patent classification terms. For this, a concordance table is essential, www.uspto.gov, operated by the USPTO, operates the service of conversion between the USPC and the IPC, and discloses a concordance table. Furthermore, in regard to Japanese patent information, a concordance table related to themes of the FT and the IPC is being sold. In the case in which there is a concordance table, speed analysis is performed with respect to the USPC, a USPC term is modified into an IPC term or some other patent classification term with reference to the concordance table for a specific applicant and/or a specific technical field meeting the preset criterion, and the modified patent classification term is included in a search query and then a search for speed analysis and the analysis of search results are performed for a second country, other than the U.S. (of course, when

there is an applicant or a key word, the translation of the applicant or key word into the second language is also included).

[268] Although the above various examples have been described with a focus on the number of patent applications, a similar method may be applied to other analysis indices. Other analysis indices may include 1) concentration, 2) a patent activity index (AI), and 3) occupation rate. The number of patent registrations and registration rate relevant to the number of patent applications may be indices, and the inventor or agent relevant to the applicant may be an index. The 'concentration' is an index of the extent to which a specific applicant concentrates on a specific field, is represented using a value obtained using "the number of all patent applications for a specific patent classification term/the number of all patent applications filed by the applicant* 100" based on a specific country and a specific year or preset period. The 'patent activity index' is obtained using "{(the number of all patent applications filed by the applicant for a specific patent classification term/the number of all patent applications filed by the applicant)/(the number of all patent applications for a specific patent classification term/the number of all patent applications)* 100} based on a specific country and a specific year or preset period.

[269] Meanwhile, the following item must be considered when the number is counted in speed analysis or the like. When a single patent application is associated with two or more 1) patent classification terms, 2) applicants, 3) inventors, or 4) agents, how to handle this problem is an issue. This is related to policies, and there may be three methods, that is, a method of performing individual counting, a method of performing equal division according to the relevant number, and a third method. 'Individual counting' is a method of assigning each applicant a weight of 1' when the number of applicants is two, and 'equal division' is a method of performing counting such that each applicant has a weight of 0.5 in the case. In equal division, a value having one or more digits to the right of the decimal point may be obtained as the number information.

[270] Meanwhile, from the point of view of the modified patent classification term inclusion search, which is the core of the present invention, when 2 or more patent classification terms appear, the following method, rather than the above two methods, may be considered. Two or more patent classification terms appearing in a patent document may be grouped based on a patent classification term included in a search query. For example, when a specific patent document is associated with five patent classification terms, the five patent classification terms may be divided into a first group, including a patent classification term, which is itself included in the search query, and lower patent classification terms, and a second group, including higher patent classification terms above the patent classification term included in the search

query and patent classification terms relevant to separate, independent nodes/layers. In this case, a method of assigning a different weight depending on the number or distribution rate of component patent classification terms of each of the first and second groups may be employed. For example, in the case in which a patent document is associated with H04B 7/005, H04B 7/01, H04B 7/145, H04B 7/204 and H04B 5/00, there are H04B 7/145 and H04B 7/204 below H04B 7/14 if H04B 7/14 is included in a current search query (there is a search query at the time of analysis, and a patent classification term is included in the search query). Therefore, according to the modified patent classification term inclusion search method, the patent document including H04B 7/145 and H04B 7/204 is included in search results, and the patent document may be assigned 2/5, rather than 0, based on the search query reference H04B 7/14. In such a case, when a typical method, other than the modified patent classification term inclusion search method, is used, the patent document is not included in search results if H04B 7/14 is included in a search query.

[271] The modified patent classification term inclusion search method can be usefully applied not only to the analysis but also to monitoring or watching service. The monitoring service refers to a service of registering a specific search query, and transmitting new patent information to a monitoring service user when new patent information relevant to the search query is found. In the present invention, the monitoring is performed by a patent information monitoring engine 300. The patent information monitoring engine 300 functions to receive the registration of a search query to be monitored, query the search engine 110 or DBMS 170 using the search query at preset time intervals (every day, every week, every month, every quarter and/or every year), and, if a new patent document occurs, provide notification of information about the occurrence of the new patent document, that is, the fact, the document, or information about the fact or document, to a monitoring service user. In this case, one or more of a method of transmitting monitoring results to a receiver via electronic mail or other message transmission means and a method of providing notification to the monitoring service user when he or she accesses the monitoring service may be employed as the method of providing notification to the service user.

[272] In order to perform monitoring service, the steps of receiving the registration of a search query to be monitored; querying the search engine 110 or DBMS 170 using the search query at preset time intervals (since a query aims to search for one or more new documents, a time range (for example, the time range is defined as the previous 6 days including a publication date when the search query is based on the publication date) is generally defined in a search query), and transmitting one or more patent documents to a monitoring service user if patent documents are found may be performed. Here, in the case in which a patent classification term is included in a search query, the

monitoring service may be provided using the modified patent classification term inclusion search method. In the monitoring service using the modified patent classification term inclusion search method of the present invention, the problem in which a patent document is excluded from monitoring results in the case in which a patent classification term included in a search query does not exist in the patent document although the patent document logically and hierarchically falls under the patent classification term included in the search query can be prevented.