Strategic Decision Support and Identify New Customer Segments and Mitigate Risk Management for Loan Schemes Using Data Mining

Abstract

Banking sector is fast growing industry in Sri Lanka with the novel technological achievements, so there is a massive competition over the major banks which will try to attract potential customers to provide a better customer service. Like organizations, banks also face the problems on getting quality information in time to make critical decisions. Depend on the quality and the way service provide affect badly to increase number of customers and retain of existing customers. However banks achieve this task using information communication technology (ICT) depend on the information systems as well as technical competencies. Banks provide services through various channels like Automatic Teller Machines (ATM), debit cards, credit cards, internet banking etc. Within the masses of information in the bank databases lies hidden information of strategic importance. Data mining is the important process in finding the particular patterns and relationships that can help their business. Therefore data mining and decision making is emerging in banking sector in worldwide these days. Customer data are the raw material that must be captured, integrated and effectively analysed in order to achieve the goal of profiling customers. The key enabler of any particular categorisation or segmentation strategy is customer data. Effective segmentation requires both static customer profile data which includes demographic, job status, preferences and customer behaviour data which are product benefits, rewards and so on. Once these data are integrated banks can be interrogated to discover groups of customers sharing the same characteristics and needs. Identifying customer segmentation is the process of partitioning the heterogeneous customers into separate and distinct homogeneous segments. In this research, identify more potential customer segments when launching new banking products for different market segments and services such as health care programmes, donations and predict who are the best potential customers having good relationship with the bank and give alarm there is high probability a specific customers will not pay on time. The methodology used was the knowledge discovery in databases (KDD) process. In KDD process data mining technique was used as clustering. Clustering was for finding groups of similar instances in a dataset. Clustering used to evaluation based on likelihood if clustering schemes produces a probability distribution within the selected data and the KDD process was used as the framework that guided the entire process.