CLASSIFICATION RULE GENERATION PROGRAM, CLASSIFICATION RULE GENERATION METHOD, AND CLASSIFICATION RULE GENERATION DEVICE

摘要

A classification device extracts a plurality of connected character strings in which words obtained by segmenting text data in units of words are connected. For each of the plurality of connected character strings, the classification device refers to a storage part storing classification categories wherein a plurality of instances of text data including said text data are classified, and computes distribution of categories into which the text data from among the plurality of instances of text data that includes the relevant connected character string is classified. On the basis of said distribution, the classification device selects characteristic terms from the plurality of connected character strings, and generates a classification rule in which the classification categories into which the text data is classified are associated with the characteristic terms.