Approximate collation device, approximate collation method, program, and recording medium

Abstract

Keywords can be extracted at high speed even if wrongly written words, skipped-character words, expressions a little different from expected keywords are contained in an inputted character string and the number of keywords is large. A skip dictionary creating unit (10) creates a skip dictionary containing the keywords to be extracted and listed in the keyword dictionary and a set of deletion keywords made by deleting at least one character in every character position from each keyword and stores the skip dictionary in a skip dictionary storage unit (20). A keyword extracting unit (30) collates an inputted character string with the words in the skip dictionary, extracts a preliminarily-given keyword and keywords approximate to the keyword from the inputted character string, and outputs them along with the appearance positions.

Description

technical field [0001] The present invention relates to a technique of comparing an input character string with a predetermined keyword in order to extract a predetermined keyword from a text (input character string) described in natural language, and outputting a matching keyword and its appearance position. Background technique [0002] <keyword extraction> [0003] Keyword extraction is a task of extracting keywords listed in advance such as a dictionary from an input character string described in a natural language. [0004] For example, consider starting from figure 1 The keywords related to the Olympic Games are extracted from the input string shown. In this case, by performing figure 2 As shown, the keyword to be extracted is extracted by comparing whether each keyword in the keyword set (hereinafter referred to as keyword dictionary) manually included in the table is included in the above-mentioned input character string. . [0005] However, there are the...

AI Technical Summary

Problems solved by technology

[0006] 1. Because the input string is manually described (input), it often contains typos and omissions, or uses slightly different expressions from the expected keywords;

[0007] 2. If the number of keywords increases, quick extraction becomes difficult, and it takes time to process a large number of input strings

[0034] However, since the AC method can only extract character strings that are exactly the same as the keywords, there are problems that typos and omissions cannot be extracted like the trie structure, and keywords with slightly different expressions are used.

Images