<論文>決定木アルゴリズムによるウェブページの分類法の提案 (<特集>情報メディア)

概要

論文の詳細を見る
An enormous web pages are published over the Internet and are going to be used for market researches. At the current status of web pages, there is no standard format to classify pages though particular types of pages are often required form a marketing viewpoint. This paper addresses the issues of web page classification problem, whose goal is to find the best logic formula that identifies a set of web pages from the whole web page database. In the proposed algorithm, web pages are characterized with an array of boolean variable that tells whether or not a keyword is included by the page. Information gain is used to evaluate the best classifier for the given web page samples. Finally, the resulting logic formula is used to make a survey of population in terms of favorite sports.
東海大学の論文
2002-09-30