Prediction The Amount of Collected Waste using Deep Neural Network Maimunah
Universitas Muhammadiyah Magelang
Abstract
The public health of residents has been affected by the increase in unhealthy waste management in cities of developing countries. Waste management has received extensive attention from the city government. The plan and design of a sustainable municipal waste management system requires accurate predictions of the solid waste generated and collected by municipal authorities in landfills. The TPA in Magelang City receives waste collection from several waste suppliers which is weighed and recorded every day. Based on the existing data, the addition of the amount of waste does not decrease, so a method is needed to handle the amount of waste so that it can be managed properly. The main objective of this study is to develop an accurate prediction model for the amount of waste in landfills. Machine learning algorithm is applied to build the model. A data pre-processing and integration framework was developed in Python computing software to generate datasets with sufficient quantity and quality of data for modelling. The results show that the neural network algorithm has been successfully used to produce a predictive model for the amount of waste in landfills, both on a daily basis and as a whole. Prediction performance was measured using RMSE and MAPE.
