Saderat Bank customer classification using decision tree Based on customer value

• Main Authors: Hosein Beyorani, Mahram Azimi
• Format: Article
• Language: fas
• Published: Islamic Azad University, Tabriz Branch 2015-08-01
• Series: مدیریت بهره وری
• Subjects: data mining; decision tree; customer relationship management; customer value; customer lifetime value
• Online Access: http://jpm.iaut.ac.ir/article_519531_97d6036fd4a3a926b3bbf62a1a5a4734.pdf
• ISSN: 2716-9979; 2476-7298

Customer satisfaction in today's business environment has gained importance. Many companies to increase profits and customer satisfaction are focused on customer value. Customer relationship Management (CRM) is tools to enhance customer relationships has emerged as the principal competing companies. Successful customer relationship management in companies start from identifying customer value beacuse Customer value provides important information for the development and management. Techniques such as data mining has led to the development of customer relationship management in new competition areas so that companies can be profitable in business competition. Through data mining -the discovery of hidden knowledge database- organizations can identify valuable customer and predict their future behavior and take useful and knowledge-based decisions. The purpose of this research is to gain the effective features to select valuable customer that can classify customers based on population characteristics and other variables relating to transactions in classes very low profit, low profit, high profit and very high profit. In this study the influence of demographic characteristics including age, education and job level also affect the degree of branch, bank branch location and number of transactions on customer value will be checked. Dependent variable in this research is customer value that is classified into four categories. Statistical population is the all customers have an active checking account with Bank Saderat Iran in Tabriz city. To review the case, the CHAID decision tree data mining algorithms were used.Results showed that the variables age, education level of customer and bank branches have no significant effect on customer value and the number of customer transactions with bank has most effective in identifying the class of customer.