INFORMATION EXTRACTION SYSTEM, INFORMATION EXTRACTION METHOD, INFORMATION EXTRACTION PROGRAM, AND INFORMATION SERVICE SYSTEM

摘要

Words and phrases of the same type can be extracted from a plurality of documents of various forms. A storage device (400) stores the documents of various forms. A pattern candidate creating means (11) receives a list of input words selected as samples from words and phrases which are to be included in a dictionary. The pattern candidate creating means (11) selects one document, determines character strings before and after the input word in the document as pattern candidates, and stores them as pattern candidates (16). The pattern candidate creating means (11) performs this processing for each document. A word and phrase candidate creating means (12) extracts the words and phrases sandwiched between the patterns included in the pattern candidate (16) as word and phrase candidates which are to be outputted and stores them as a word and phrase candidate (17). A word and phrase selecting means (13) outputs a word and phrase candidate satisfying a predetermined condition out of the word and phrase candidates included in the word and phrase candidate (17) as an output word to an output device (300).