A Data Driven Framework for Customer Segmentation and Prospective Customer Prediction in Direct Marketing

Data Mining(DM)/Business Intelligence (BI) plays significant role in discovering knowledge from various data repositories and supporting better decision making. Depending on the business area, application and requirement analysis various data mining techniques are employed leading to growth in business sectors. In Banking Industry Data Mining techniques are widely used in Credit Analysis, Customer Relationship Management, Customer Segmentation, Marketing, Fraudulent Transactions, Risk Management and many more. A data driven framework for predicting prospective customers that are to be targeted for bank direct marketing is proposed in this paper. An approach of contacting all customers in the repository for new business plan is replaced by targeted groups of customer segments that yield to prospective business. The predictive model built upon demographic features of customer in the bank database will select the prospective customers upon which the bank can apply direct marketing. Two classification models Naïve Bayes and Logistic Regression are studied using bank dataset from UCI machine learning repository and their accuracies are compared. The experiments are done in Weka machine learning simulator. Logistic regression showed better accuracy with reduced and enhanced feature set