摘要 |
PROBLEM TO BE SOLVED: To offer a document cluster processing apparatus, a document cluster processing method and a program which can give an evaluation value suitable for human feeling, the evaluation value showing a merit of a coherence of a document cluster. SOLUTION: The document cluster processing apparatus has a split means to divide a document into a word or a character string, a document analysis means to give "1" to the word or the character string concerned if each of all words or character strings that appear in the document cluster appears in all documents contained in the document cluster, and to give "0" to the word or the character string concerned if it does not appear; a distribution computing means which computes the distribution in the document in the cluster about each word or character string; and an evaluation value computing means which computes an average value of the distribution of all words or character strings which appear in the document cluster, as an evaluation value of the document cluster concerned. COPYRIGHT: (C)2008,JPO&INPIT
|