Improving the Web text content by extracting significant pages into a Web site

Web systems have reached a very important role in today's business world. Every day organizations fight to keep their present clients and to gain new ones. In order to accomplish this goal it is very important to make precise changes in the Web site content. However, the development of these improvements is a complex and specialized task because of the nature of the Web data itself. We propose a novel approach to successfully make changes to improve the Web site content using text mining. We use a self organizing feature map (SOFM) to find the most relevant text content, and then we propose a reverse clustering analysis in order to extract the most significant pages of the whole Web site. The effectiveness of this method was experimentally tested in a real Web site.