Research And Application Of Customer Analysis Models Based On Data Mining

Data mining tries to find knowledges that people are interested in, and these knowledges are informations that are connotative, unknown beforehand and useful. As far as customer analysis is concerned, this paper adopted the method of data mining, aiming at the problems of tradional methods of customer analysis, such as low efficiency and low exactitude. Meanwhile, the current method of dataming only aims at the material problems to set models, in order to settle commercial problems. So they can not be used in other aspects. Therefore, in order to settle customer analysis problems ,this paper adopted a new data mining system which can be used in many different problems.From analyzing the commercial problems, this paper constructed a data warehouse CDW, which surrounds the client and can be extended. Meanwhile, this paper designed customer analysis models such as classification, churn, cross sales and customer respondense on the CDW, and realized them on the data mining tool DMiner. DMiner is a visual analysis platform which is designed for analysis personnel to analyse data by means of technologies of data mining and statistic analysis, and the author has participated in the exploitation.This paper discussed particularly on the steps of designing models, choosing of data mining arithmetics and parameters and the laying out of the result, then we can choose the best one as the finally model. Furthermore, on the base of the classifying arithmetic SLIQ that will be used in the models,this paper improved it ,which is called DM_Tree arithmetic,and applied it in the data mining models.On the base of the new style of data mining system's structure and the customer analysis models which were discussed above, this paper applied these models in commercial banks and realizing the data mining module of commercial banks'customer analysis, accomplished data mining function such as classification and churn,which has obtained good effects.