Section-Based Filtering and Its Application to SearchFilter

Xie Xing,Liu Wenyin,HongJiang Zhang,Yudong Yang

2001-01-01

Abstract:Demands for automatic filtering of restrictive contents on the WWW are growing continuously as the accessibility of the web is rapidly expanding. Most information filtering technologies aim at site-filtering or page-filtering, which are often inconvenient since other pages of the same site or other sections in the same page may still be appropriate for the reviewer. In this paper, a new framework of section-based filtering is proposed, which can be used to filter out inappropriate sections in a web page and has been applied to the search results returned by common search engines. Based on visual similarities, a search result page can be automatically parsed into sections by our system, and then each section will be filtered by a SVMbased approach which can be combined with additional user-defined rules. Experimental results show that our approach can achieve a high recall in filtering the search results.