Provide a Framework for Predicting Customer Behavior and Pattern and Determining the Most Important Criteria Using Customer Relationship Management and Data Mining Approaches (Case Study: Mellat Bank)

Banks are one of organizations that are increasingly under pressure today. If banks lose their customers, they will find their resources in danger and they may lose their organizational life. Therefore, banks are always trying to discover the secret of customer churn clearly, so that they can avoid it. Thus, this subject has been dealt with in the research. The study is applied research in terms of objective and exploratory research in terms of method of investigation. Statistical population of the research included some of actual customers of bank in Arak City whom data existed in the Bank. The numbers of these customers were 149 people. The tool to collect data was customer database. Accordingly, fifteen traits were selected for predicting customer churn. According to the results, C&R Decision Tree Algorithm could predict customer churn better than other algorithms. In this regard, it is suggested that the rules will be applied at the instructions of marketing and customer retention. According to the results, five important features for predicting customer churn are as following which banks should pay special attention to these features: job, the rate of branch, education, the averagebalance, and type of investment.

Keywords:

Customer Relationship Management (CRM) , Data Mining , Classification Algorithms , Association Rules Algorithms