Predicting Household Savings Behaviour for Financial Sustainability: Leveraging Machine Learning to Foster Economic Resilience

Abstract

Household Saving Behaviour is the base for Financial Sustainability that promotes long-term development and economic resilience. According to the Indian context, the ancient book “Arthashastra” stresses the management of resources, advocating both saving and acquisition of money as a moral and practical goal. It also stresses careful financial management, promoting the allocation of revenue for various family activities. The ancient Tamil scripture “Thirukkural” emphasizes saving as a road to manage resources wisely, supports the notion of financial resilience that leads to sustainable progress. This research tries to connect ancient concepts with new technology techniques by predicting household Saving Behaviour using different machine learning algorithms. To predict Saving Behaviour, several characteristics such as demographic and socio-economic, cultural, and civilization aspects have been incorporated to build the best-fit model. Advanced ML models such as Support Vector Machine, Gradient Boosting Classifier, Random Forest Classifier, and Naïve Bayes and others are used to predict Saving behaviour. Evaluation of each model is done using several measures such as accuracy scores, precision, recall and F1 – score to find the best-fit model. By comparing the accuracy ratings of each model, the most accurate model to predict Saving behaviour is selected.