Applied Research, Data Mining In Crm

Customer Relationship Management(CRM) is a kind of new management system aiming at improving the relationship between corporations and their customers. It enables corporation to attract more and more customers through offerring kind and characteristic services. Having use traditional MIS for many years, coporations gathered and stored mass of data about customers. A great deal of important information is concealed under those datasets. Datamining define models which describes customer classification or prediction of customer's behaviors. Those models help corporations gain useful business information from those data, such kind of information supports corporations to make decisions. Begin with understanding the needs from CRM, this dissertation introduces datamining theroies and techniques at first, then emphasis is placed on the theme how to perform customer segmentation. To abtain the target, works as follows were finished:Firstly, by evaluating problems that too many nodes may be induced by classical SLIQ algorithm even after decision tree pruning, an improved algorithm is proposed. By specifying a parameter of nodes number decision tree at most included, users can get a rough view of the patterns in their data. In order to validate the improved algorithm, a software demo of modeling decision tree is designed and programmed, customer segmentation is executed by this demo according to customers' information. The demo offer functions as: visualization of datasets and decision tree model, extracting rules from model, analysizing differernt models by adjusting input parameters for building decision tree .Finally, a decision tree model is described by using Predictive Model Markup Language(PMML) document, those kind of documents can be shared amoung differernt dataming tools.