A framework for Web Usage Mining based on Multi-Agent and Expert System An application to Web Server log files
A framework for Web Usage Mining based on Multi-Agent and Expert System An application to Web Server log files

Summary/Abstract: In this paper we introduce a Web Usage Mining (WUM) system, which transparently and dynamically generates content (links) and structure (layout) of the web portal for anonymous visitor. To this purpose we make use of the two software concepts: Multi-Agent and Expert System simultaneously. Multi-Agent system is responsible for discovering knowledge from Web Server log files which requires data preprocessing, data mining, content and structure generating. On the other hand, Expert System supports the software agents by means of cognitive and heuristic knowledge provided and verified by a human expert. The main innovation proposed here is a novel approach that can be used to manage large Web portals.

• Details
• Contents