Forecasting Waste Generation with Increment Linear Regression Technique: A Case Study of SIMASKOT Application

This research aims to develop a prediction system for urban waste generation using the Incremental Linear Regression method on SIMASKOT. This method is applied to deal with the limitations of historical data, where the prediction results from the previous year are used as training data to predict the next year. The problem faced is the lack of sufficient data to create accurate and reliable prediction models in the long term. The purpose of this study is to improve the accuracy of waste generation prediction using an incremental regression approach. The experimental methodology involves the use of waste generation data from several waste categories during the period 2019 to 2022, which is then used to predict data until 2026. Model evaluation was carried out using the metrics Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R². The results show that this incremental prediction model is able to provide more accurate predictions than conventional models, especially for more volatile waste categories such as wood twigs and metals. The conclusion of this study shows that the Incremental Linear Regression technique is effective to be used in waste generation prediction, and can be integrated in long-term prediction-based monitoring applications.