Prediction of Term Deposit in Bank: Using Logistic Model

Abstract

Telemarketing remaining to be one of the most popular direct marketing techniques, which is urged to make accurate predictions. It is critical for banks to have a mature machine learning approach for marketing forecasting problems. Through the analysis of the time deposit telemarketing data of Portuguese commercial banks on Kaggle.com, the bank can more accurately locate the target customers, so as to improve the efficiency of increasing the amount of time deposit business. Firstly, the data are preprocessed, coded and shuffled to divide the test set and training set. Secondly, logistic regression is carried out to eliminate irrelevant independent variables through correlation analysis, which makes the prediction more accurate. The results of logistic regression were compared with the results of decision tree and retrograde comparison. Accuracy and Area Under Curve (AUC) were used as evaluation indexes. The results demonstrate the effectiveness of the logistic regression model and the elimination of irrelevant variables. This method can be applied to the actual marketing problems of banks to accurately find target customers and improve the accuracy of marketing.