Abstract
In this high competition age, a company has to continuously keep an eye on the latest information in order to hold the upper hand of the industry. The company may have to find the information on the mass media or on the market. They can even find useful information in their own database. The task of mining unseen information and then transforming it into the competitive strategy is essential in the data mining area. Customer relationship management system is one of the most popular data mining applications. In this study, we analyze a subsystem of a 3C retailer’s CRM System ---an eCard recommendation system. At the same time, we propose an architecture for incremental data classification. We then apply this technique to the eCard recommendation system to see whether it would perform better than the existing ones. Experimental results show that the classifier built according to the proposed method has acceptable error rate compared with the existing classifiers. Moreover, it can generate a set of rules which provide some high level semantic description about the data.