Comparison of Artificial Neural Network (ANN) and Multiple Regression Analysis for Predicting the Amount of Solid Waste Generation in a Tourist and Tropical Area—Langkawi Island

Prediction of the accurate amount of solid waste is difficult work because several parameters affect it. There is a high degree of fluctuation in the prediction of amount of solid waste generation. Therefore, applying neural network as intelligent system can be a good option. In a tourist area such as Langkawi Island, protection of the area and pollution control are important issues; also it is significant for planning managers to obtain accurate forecasting of the quantities of solid waste generated. In this paper, weekly data of solid waste generation, types of trucks and their trips, number of personnel in per trips (entrance to landfill) during 2004-2009 have been used as variables, through feed forward back propagation, in the testing and training processes. In the last step, the best model for forecasting waste generation in Langkawi Island was selected based on mean absolute error, mean absolute relative error, correlation coefficient and threshold statistics. Validation of model is calculated for different hidden layer. Comparison between final result of ANN and Multiple Regression Analysis (MRA) showed the result of ANN is better than MRA, which suggests that ANN is a better modeling tool. Therefore, in terms of predictive accuracy test, the ANN has a higher accuracy than regression analysis. Keywords—Prediction of Solid Waste Generation, Langkawi Island, Artificial Neural Network, Multiple Regression Analysis, comparison