To achieve retrieval with a natural language by processing simpler than morphological analysis.
A first information processing apparatus applies morphological analysis to each document included in a predetermined document set, generates a set of extracted morphemes as an appearance morphological set, and generates retrieval index information showing the relationship between documents in the document set and morphemes in the appearance morphological set. A second information processing apparatus for storing the document set, the appearance morphological set, and the retrieval index information receives input of a retrieval query to the document set, and extracts a partial character string of the retrieval query as a morphological candidate. The second information processing apparatus determines whether each morphological candidate matches with morphemes in the appearance morphological set, and calculates similarity of each document to the retrieval query based on the morphological candidate determined to match and the retrieval index information. The second information processing apparatus presents a document similar to the retrieval query based on the similarity.
COPYRIGHT: (C)2011,JPO&INPIT
Akira Karasuya
JP2005242455A | ||||
JP2003006216A | ||||
JP2004005749A | ||||
JP11175541A | ||||
JP2002073680A | ||||
JP6103309A | ||||
JP10003481A | ||||
JP7325834A | ||||
JP9231234A | ||||
JP64031227A |
virtue Tamio Ei