PROGRAM, METHOD, APPARATUS AND SERVER GENERATING CO-OCCURRENCE PATTERN FOR DETECTING NEAR-SYNONYM

摘要

PROBLEM TO BE SOLVED: To provide a program, a method, an apparatus and a server capable of generating a co-occurrence pattern for excluding a pattern which has high generality and a high co-occurrence with a seed word set to avoid near-synonym candidates having weak relation with seed words from being retrieved.;SOLUTION: The apparatus includes: seed sentence search means that searches for seed sentences which include several seed words s from a large amount of sentences; feature word detection means that calculates co-occurrence frequency C(w, s) of appearance of the seed words s and words w by using every seed sentences, and calculates an evaluation value Score(w) based on the co-occurrence frequency common to every seed words s with respect to each word w to detect a common feature word w which has the evaluation value Score(w) larger than a predetermined threshold value; word string detection means that detects word strings of a predetermined length in which both of the seed words s and the common feature word w appear; and co-occurrence pattern generation means that generates a co-occurrence pattern in which a part of the seed words is replaced with a wild card with respect to each of the word strings.;COPYRIGHT: (C)2015,JPO&INPIT