Inter-word relevance calculation device, inter-word relevance calculation method, inter-word relevance calculation program, and recording medium recording the program

摘要

PPROBLEM TO BE SOLVED: To provide an inter-word relation degree calculation unit having high accuracy. PSOLUTION: The unit includes a word dictionary 102 storing combination of sets of words, information of parts of speech, and semantic information of the words; a morphological analysis means 101 for performing morphological analysis of text by referring to the dictionary 102; from the analysis result of the morphological analysis means 101, a vector generation means 103 for calculating an event frequency of coincident occurrence of a pair of an arbitrary word and arbitrary semantic information in the word dictionary 102, which is counted over the overall text, and to each word, generating a vector of which each coordinate value is the above frequency calculated for the pair of the word and the semantic information made correspondence to the above coordinate; a relative frequency degree calculation means 104 for converting each coordinate value of the vector to a relative frequency; and to a word pair, a Kullback-Leibler distance calculation means 105 for calculating a Kullback-Leibler distance between the vector corresponding to one word and the vector corresponding to the other word. PCOPYRIGHT: (C)2008,JPO&INPIT