Personalization for Web-based Book Shop System using Hybrid Data Mining Approaches

Abstract

Personalization is a new system development approach for designing information systems that change configurations based on each user's needs and preferences. If it is possible to recommend products to customers’ liking at the time they are visiting the specific web site, it would reduce the hassle customers experience in searching for products from a large information base. An effective personalization technique has to be customized to meet the specific needs of every particular domain and deliver quality recommendations. This paper presents a hybrid personalization system for web-based book-shop system. It combines the Bayesian classification method with association rule mining to model individual customer’s behavior. While Bayesian classifier is for effective customer profiles, Association rule presents item to item association based on user transactions.