Application of Machine Learning and Deep Learning to Predict Financial Product Subscriptions Based on Customer Features

Abstract

The financial industry faces challenges in predicting consumer behaviour, especially in forecasting decisions related to subscribing to financial products like term deposits. This study applies machine learning and deep learning to predict subscriptions based on demographic and behavioural data from the Bank Marketing dataset from the UCI Machine Learning Repository. The models tested include Support Vector Machine (SVM), Logistic Regression (LR), Random Forest (RF), Bidirectional Gated Recurrent Unit (BiGRU), and Bidirectional Long Short-Term Memory (BiLSTM). Model performance is evaluated using metrics such as accuracy, precision, recall, F1-score, and confusion matrix. Results show that BiGRU achieves the highest accuracy of 92.52%, outperforming other models, with SVM and BiLSTM also showing strong performance. However, all models still face limitations in detecting subscribing customers, as evidenced by the high false negative rate. These findings highlight the potential of machine learning and deep learning to support data-driven decision-making in financial marketing, despite limitations such as the use of a single data source and the lack of consideration for external factors affecting customer decisions.