A Decision Support System for Predicting Bank Target Clients: Comparison between Decision Tree and Artificial Neural Network

Abstract

Nowadays, many banks face a high cost of operation for finding bank clients via telemarketing.  A possible solution to solve this problem is finding an effective technology to support work processes and decision making. The effective technology to support bank operation helps banking business continuously growing, survive and make a profit. Therefore, this work proposed a model for prediction an opportunity of the customer to be a target customer of a new bank campaign to reduce the operation cost of finding a new customer. The results from the model were used to develop a decision support system program for predicting target bank clients to promote marketing campaigns via phone call. The system focuses on to predict a possible rate of the customer to subscribe to a bank deposit. The proposed model has compared an accuracy rate between decision tree classification algorithm and artificial neural network (ANN) to predict bank target clients who subscribe bank deposit. The results showed that the decision tree algorithm to predict a bank target client which improved the effectiveness by applying a fuzzy set was high accuracy rate at 88.38%.