PROBLEM TO BE SOLVED: To properly extract advertisement pages without decreasing the precision of results obtained by extracting reputation information from Web pages and analyzing them.;SOLUTION: The Web page sorting program that makes a computer extract advertisement pages on which articles described by advertisers are put from Web pages on which articles are put on the Internet stores a list of phrases in which phrases consisting of unique expressions are registered, extracts phrases from the text information included in the Web pages, counts the number of cases in which the phases in the phase list match the extracted phrases, and extracts the advertisement pages from the Web pages based on the counts (Since it is assumed that many phrases consisting of unique expressions are included in the text information included in the advertisement pages, the Web pages that contain more phrases consisting of unique expressions than the set threshold are extracted as the advertisement pages, for example.).;COPYRIGHT: (C)2008,JPO&INPIT
展开▼