Solid waste generation forecasting by hybrid of artificial neural network and wavelet transform

Quantitative Prediction of municipal solid waste generation has an important role in the optimization and programming of municipal solid waste management system. But, this concept was companied with many problems, because of the non homogenous nature and the effect of various factors out of the control on solid waste generation. In this study, the combination of artificial neural network and wavelet transform (wavelet-neural network) is used to predict the weekly generation in Tehran, concerning complexity and dynamic municipal solid waste management system. In order to this forecasting, time series of generation of this city arranged weekly in the period of 1380 to first three months of 1385, are used. The results achieved in this research indicate the positive effect of preprocessing of input variables by the wavelet transform in prediction of weekly generation in this city so that it has led to noticeable increasing in the accuracy of model calculation. The correlation coefficient (R2) of models, in the stage of testing, has improved from 0.41 in the model of neural network to 0.91 in the model of wavelet-neural network.