PREDICTING CUSTOMER SUBSCRIPTION TO TERM DEPOSITS USING MACHINE LEARNING: A RANDOM FOREST APPROACH

Abstract

a machine learning-based predictive model to identify the likelihood of customer subscription to term deposits based on demographic, financial, and campaign data. Using a Random Forest Classifier, the model achieved an accuracy of 91% and a ROC-AUC score of 0.94. The methodology incorporates comprehensive data preprocessing, feature engineering, and model evaluation to ensure robust predictions. The insights derived can assist banks in optimizing marketing strategies and resource allocation.