How does Machine Learning Predict the Success of Bank Telemarketing?

This paper intends to understand the antecedents in the success of bank telemarketing prediction modeling better and seeks to evaluate the predictive performance of machine learning. For this, we are using a machine learning technique on the dataset of direct marketing campaigns of a Portuguese banking institution which is obtained from the UC Irvine Machine Learning Repository. Based on these results, first, among all variables, age, balance, loan, day, duration, campaign, pdays, and poutcome influence the success of bank telemarketing, while job, martial, education, default, housing, contact, month, and previous have no significance. Second, for the full model, the accuracy rate is 0.784, which implies that the error rate is 0.216. Among the patients who predicted not to have the success of bank telemarketing, the accuracy that would not have the success of bank telemarketing was 75.63%, and the accuracy that had the success of bank telemarketing was 82.61% among the patients predicted to have the success of bank telemarketing. This study extends the existing literature by empirically examining the combined impact of the variables on the success of bank telemarketing modeling.