METHOD AND DEVICE FOR GENERATING DICTIONARY, AND RECORDING MEDIUM RECORDED WITH DICTIONARY GENERATING PROGRAM

摘要

PROBLEM TO BE SOLVED: To generate a meaning category dictionary showing which kind of words are a representative vocabulary in each meaning category in accordance with an application field to be handled. SOLUTION: A corpus analyzing part 11 reads two or more kinds of corpuses from a corpus storage part 10 and obtains information for identifying a main noun appearing in each corpus from the respective corpuses. A main noun extracting part 12 extracts the main nouns (a common main noun) appearing in every corpus based on the analysis result and calculates the appearance frequency of each common main noun at each corpus. A score information imparting part 13 imparts a score to the common main nouns in accordance with weight values which are decided by the order of the meaning category of a word meaning in each extracted common main noun and also in accordance with the appearance frequency of the common main nouns in the respective corpuses. A register noun selecting part 16 rearranges the common main nouns in order from the one with the high scores which are imparted at each meaning category, selects a fixed number of common main nouns from the high order one and stores them in a meaning category record storage part 17.