Food waste behaviour prediction and policy recommender system: data-driven machine learning approaches

The rising economic, environmental, and social consequences of food waste necessitate comprehensive studies to identify the primary drivers of household food waste (HFW) and to develop effective strategies for reducing it. This study predicts Food Waste Behavior (FWB) using a two-phase approach. In the first phase, an extensive literature review, complemented by expert opinions, is conducted to identify the most influential factors contributing to HFW. A Structural Equation Model (SEM) is then developed and validated using data obtained from a customized survey of 300 Omani households to establish the theoretical relationships among these factors. Next, IF–THEN rules derived from expert input through a structured Delphi process were used to assign FWB levels to all possible combinations of identified indicators via a Full Factorial Design (FFD). This dataset was then used to train several machine learning (ML) models—including Decision Tree (DT), Random Forest (RF), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), Light Gradient Boosting Machine (LightGBM), and Adaptive Neuro-Fuzzy Inference System (ANFIS)—to predict FWB in households. Among these models, MLP and SVM exhibited the best performance, achieving 99% classification accuracy.