C4-2: Combining Link and Contents in Clustering Web Search Results to Improve Information Interpretation

Yitong Wang,Masaru Kitsuregawa

2002-01-01

Abstract:With information proliferate on the web, it is far beyond human’s ability to digest this huge, heterogeneous information, e.g. locating related resources as well as providing accordingly information interpretation. While web search engine could retrieve information on the Web for a specific topic, users have to step a long ordered list in order to locate the needed information, which is often tedious and frustrating. In this paper, we investigate how to combine link and contents analysis in clustering web search results to improve information interpretation for a specific topic. By filtering some irrelevant pages, the proposed approach clusters high quality pages in web search results into semantically meaningful groups with additional tagging keywords to facilitate users accessing and understanding. We especially study the contribution of link and contents to clustering procedure. Preliminary experiments and evaluations are conducted to investigate its effectiveness.