An extension of the theory of planned behavior with machine learning to predict household food waste in India

Abstract

Food wastage is a complex social, economic, and environmental issue worldwide. Approximately one-third of edible food products are wasted globally. There is an urgent need to raise awareness about food wastage. The Theory of Planned Behavior (TPB), which encompasses attitude, subjective norm, and perceived behavioral control, has been extensively used to study food-related behaviors. In this study, we propose an extension of TPB by adding three attributes: self-responsibility, eco-consciousness, and food and home management, aiming to develop a more comprehensive framework for predicting food wastage behavior. A cross-sectional survey was conducted among individuals of different age groups, measuring the impact of various attributes on food wastage. A total of 341 responses were collected across India. The data were analyzed using statistical methods and machine learning (ML) algorithms such as Naive Bayes (NB), Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF). The prediction accuracy of these ML classifiers for TPB and its extension was compared. Additionally, explainable artificial intelligence (XAI) techniques were employed to assess feature importance and identify significant attributes influencing food wastage. Results indicated that the prediction accuracy of all ML classifiers improved by 12% after incorporating self-responsibility, eco-consciousness, and food and home management attributes, thereby validating the extension of TPB. The integration of XAI highlighted key attributes associated with food wastage and increased the transparency of complex predictive models.