Title:
【発明の名称】重要文検索装置
Document Type and Number:
Japanese Patent JP2788283
Kind Code:
B2
Abstract:
PURPOSE:To retrieve the sentences included in a document containing the desired contents with high efficiency by using the information on the structure, the attribute, etc., of the sentences included in the text having a part to be collated with a given key word. CONSTITUTION:An input/output part 1 inputs the key word, and the attribute of a desired sentence to be retrieved, the relation of the sentence with the document structure, etc., and also outputs this retrieved sentence. A key word collator 2 collates the given key word with a character string in the text of a document and presents the collatable sentences. A sentence structure/attribute analyzer 3 performs the structural analysis of a single sentence and the analysis of the sentence attribute. A document structure analyzer 4 recognizes a rough logical structure of an introduction, a conclusion, etc., proper to the documents, the titles, and the physical structures of documents. Thus it is possible to correctly and effectively extract the sentences containing the information desired by users out of the documents by using the attributes and the structures of sentences.
Inventors:
TSUTSUMITAKE HIDEYUKI
Application Number:
JP9750989A
Publication Date:
August 20, 1998
Filing Date:
April 19, 1989
Export Citation:
Assignee:
TOSHIBA KK
International Classes:
G06F17/30; (IPC1-7): G06F17/30
Domestic Patent References:
| JP6329882A |
Attorney, Agent or Firm:
Hideaki Tokawa (1 outside)