Customer Profiling using Classification Approach for Bank Telemarketing

Abstract

Telemarketing is a type of direct marketing where a salesperson contacts the customers to sell products or services over the phone. The database of prospective customers comes from direct marketing database. It is important for the company to predict the set of customers with highest probability to accept the sales or offer based on their personal characteristics or behavior during shopping. Recently, companies have started to resort to data mining approaches for customer profiling. This project focuses on helping banks to increase the accuracy of their customer profiling through classification as well as identifying a group of customers who have a high probability to subscribe to a long term deposit. In the experiments, three classification algorithms are used, which are Naïve Bayes, Random Forest, and Decision Tree. The experiments measured accuracy percentage, precision and recall rates and showed that classification is useful for predicting customer profiles and increasing telemarketing sales.