Data Mining and User Profiling for An E-Commerce System

Many companies are now developing an online internet presence to sell or promote their products and services. The data generated by e-commerce sites is a valuable source of business knowledge but only if it is correctly analysed. Data mining web server logs is now an important application area for business strategy. We describe an e-commerce system speci cally developed for the purpose of demonstrating the advantages of data mining web server logs. The data generated by the server logs is used by a rule induction algorithm to build a pro le of the users, the pro le enables the web site to be personalized to each particular customer. The ability to rapidly respond and anticipate customer behaviour is vital to stay ahead of the competition. The e-commerce site is written in Java and uses an o -theshelf data mining package and a freely available webserver.