Mining web hyperlink data for business information: The case of telecommunications equipment companies

Authors

L. Vaughan, Faculty of Information and Media Studies, University of Western Ontario, London, ON, N6A 5B7, CanadaFollow
J. You

Document Type

Conference Proceeding

Publication Date

1-1-2005

Journal

1st IEEE International Conference on Signal-Image Technology and Internet-Based Systems, SITIS 2005

First Page

190

URL with Digital Object Identifier

https://doi.org/

Last Page

195

Abstract

Few studies have explored the possibility of mining Web hyperlink data for e-commerce or business information. This study is an attempt to f ill this gap. The project selected a group of telecommunications equipment companies and collected data on the number of links pointing to the company Websites (Minks) and the company's revenue. A significant correlation between the two variables was found, suggesting that Minks contain useful business information and can be objects for Web data mining. The project then explored the feasibility of using Web co-link data to map business competition positions. Co-link data (links pointing to a pair of company Websites) were collected and analyzed using multidimensional scaling (MDS). MDS maps correctly clustered these companies into sectors of the telecommunications equipment industry. Data collection was repeated after nine months and the results confirmed the reliability of the methodology developed in the study. © IEEE SITIS 2005.