Hybrid fuzzy machine learning model for municipal solid waste forecasting in India

Abstract Accurate forecasting of municipal solid waste (MSW) generation at the state level is essential for effective planning and policy formulation in India. This study presents a hybrid fuzzy–machine learning framework for short-term, one-step-ahead forecasting of state-level MSW generation under data uncertainty. The approach integrates fuzzy membership–based feature representation as an uncertainty-aware feature engineering step with regression-based machine learning models. Trapezoidal fuzzy membership functions are employed to represent uncertainty in historical MSW data compiled for Indian states over a five-year period, and the resulting fuzzified dataset is used to train Random Forest, Support Vector Regression, and Extreme Gradient Boosting models. Model performance is evaluated using standard five-fold cross-validation. The results indicate that incorporating fuzzy-derived membership degrees leads to a modest but consistent improvement in forecasting accuracy, with the best-performing model achieving a coefficient of determination $$(R^2)$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mo>(</mml:mo> <mml:msup> <mml:mi>R</mml:mi> <mml:mn>2</mml:mn> </mml:msup> <mml:mo>)</mml:mo> </mml:mrow> </mml:math> of 0.766 and a root mean square error of 2382.42 tonnes per day, providing an uncertainty-aware, interpretable decision-support tool for short-term state-level MSW forecasting.