Automated Identification of Web Communities for Business Intelligence Analysis
Mcl Chau, Kw Shiu, Isabelle Yee Shan Chan, H Chen · 2005
Analysts often search the Web for business intelligence using traditional search engines which provide keyword-based search. Recently, it has been suggested that the incoming links, or backlinks, of a company’s Web site can provide useful information about the company’s “Web communities”. Backlinks refer to other Web pages which have a hyperlink pointing to the company of interest and these pages form a cyber community on the Web. Analysis of these communities can provide useful signals for a company or information about its stakeholder groups, but the manual analysis process can be very time-consuming for business analysts and consultants. In this study, we report the design and evaluation of a tool called Redips that integrates automatic backlink meta-searching and text mining techniques to facilitate users in identifying such cyber communities on the Web for business intelligence purposes. The system architecture of the tool is presented and an experimental study was reported. The experiment results showed that Redips performed significantly better than two benchmark methods, namely backlink search engines and manual browsing.