UMM Logo

Universitas Muhammadiyah Malang


Free Download Journals Directory


International Journal of Management and Business Research

Customer Retention Based on the Number of Purchase: A Data Mining Approach

1 Sahar Mehregan ; 2 Reza Samizadeh
1Department of Information Technology Management, School of Management and Economics, Science and Research Branch, Islamic Azad University, Tehran, Iran 2Department of Computer Engineering, School of Computer Engineering, AL Zahra University, Tehran, Iran
Abstract :

Purpose: this study wants to find any relationship between the numbers of purchase and the income the customer brings to the company. The attempt is to find those customers who buy more than one life insurance policy and represent the signs of good payments at the same time by the help of data mining tools. 
Design/ methodology/ approach: the approach of this research is to use data mining tools based on CRISP-DM methodology. The classification is based on the K-means algorithm and prediction is applied by a proposed formula by the researcher in Excel worksheet. 
Findings: By selecting the customers who bought more than one policy, and filtering the Income bringer customers, the researcher could extract some simple rules to predict which customer belongs to which cluster. Based on the prediction, the company can change its strategies in relation to different customers. 
Originality/value: Utilizing data mining approach to classify different customers in life insurance and prediction based on the classification is a new approach amongIran insurance companies. There is not enough research and implementation in relation to the CRM and data mining in the insurance industry. Especially CRISP-DM methodology was used very hardly before in a life insurance investigation. 
 

Keywords :
Cross selling; Data mining; Prediction; Customer retention; CRISP-DM

Date Deposited : 29 Mar 2016 11:16

Last Modified : 29 Mar 2016 11:16

Official URL: http://www.ijmbr.org/article_51_13.html

Volume 2, Number 1, - 2012

Download:
Full Text Original
Abstract : pdf doc