Iraj Mahdavia - Mazandaran University of Science & Technology, Babol, Iran
Namjae Cho - The School of Business, Hanyang University, Seoul, Korea
Babak Shirazia - Iran University of Science & Technology , Tehran , Iran
Shahram Ghobadi

Internet technology enables companies to capture new customers, track their performances and online behavior, and customize communications, products, services, and price. The analysis of customers and customer interactions for electronic customer relationship management (e- CRM) can be performed by data-mining (DM), optimization methods, or combined approaches. Web mining is defined as the discovery and analysis of useful information from World Wide Web (WWW). Some of web mining techniques include analyses of user access patterns, web document clustering and classification. Most existing methods of classification are based on a model that assumes a fixed-size collection of keywords or key terms with predefined set of categories. This assumption is not realistic in large and diverse document collections such as World Wide Web. The researchers here propose a new approach to obtain category-keyword sets with unknown number of categories. On the basis of the training set of Web documents, the approach is used to classify test documents into a set of initial categories. Finally evolutionary rules are applied to these new sets of keywords and training documents to update the categorykeyword sets to realize dynamic document classification.

e-CRM; data mining; Web document clustering; neuro-fuzzy approach