Automatic web-page classification by using machine learning methods

摘要

This paper discribes automatic Web-page classification by using machine learning methods. Recently, the importance of portal site services including the search engine function on WWW is increasing. Especially, the portal site hierarchically classifying Web-pages into many topics are becoming popular such as Yahoo! service. However, the classification of Web-page into each topic category typically relies on man power, where it costs much time and care. To alleviate this problem, we propose a technique to classify Web-page automatically based on two methods for attribute generation. We apply this technique to Web-pages on Yahoo! JAPAN to construct decision trees which determine appropriate classes for each Web-page. The performance of the proposed method is evaluated in terms of error rate, recall, and precision.