WEB SEARCH SYSTEM AND WEB SEARCH METHOD

摘要

PROBLEM TO BE SOLVED: To easily search a Web page on the Internet with high accuracy.;SOLUTION: A Web search system (100) includes: Web crawl means (12) for collecting Web pages on the Internet; an information filter (12) for extracting a Web page having high similarity by calculating similarity between each Web page collected by the Web crawl means and a sample document on a first vector space created on the basis of the sample document; clustering means (32) for performing clustering the Web page extracted by the information filter on a second vector space created on the basis of the Web page extracted by the information filter; cluster identification means (16) which creates a multi-class classifier by using a clustering result as a teacher signal, and identifies to which cluster in the second vector space an unknown Web page newly collected by the Web crawl means belongs by using the multi-class classifier.;COPYRIGHT: (C)2013,JPO&INPIT