Intelligent Data Science-Driven Univariate Time Series Forecasting for Household Debris Estimation

Green material reuse stands as a beacon of sustainability in our modern world. The whole world is facing challenges such as climate change, resource depletion, and waste management. Simultaneously, the reuse of green materials provides several advantages, including maintaining ecological balance, reducing waste, conserving resources, saving energy, and achieving cost efficiency. Recognizing these advantages, researchers are focusing on predicting household garbage to facilitate the proper reuse of green materials. This paper examines various techniques for forecasting household garbage yields using machine learning models. The study selects monthly yield data from household waste as the time series and extracts seasonal information. The results of the prediction are compared with the additive seasonal ARIMA (SARIMA) model, autoregressive integrated moving average (ARIMA), and simple exponential smoothing(SES). SARIMA models offer distinct approaches that automatically detect seasonality and trends. Performance evaluation of the prediction utilizes generic metrics such as the normalized form of mean square error (NMSE), normalized form of mean absolute error (NMAE), and mean absolute percentage error (MAPE). The findings indicate that the prediction accuracy of the ARIMA model with NMSE of 1.42, NMAE of 0.09, and MAPE of 0.08 surpasses that of the SARIMA model and SES model. The insights from this study are beneficial for planning waste classification strategies developed recently. Yield estimation can greatly benefit our metropolis by enhancing domestic waste reuse opportunities.