Author(s): Reza Mohammadpour, Zahra Asaei, Aminuddin Ab. Ghani
Linked Author(s): Reza Mohammadpour Ghalati
Keywords: Neural network, rainfall, CLA, modelling, urban water.
Abstract: Prediction of rainfall is extremely important for management of water resources. In urban areas, rainfall has a great impact on traffic, sewer systems, and other human activities. In this study, Learning - Cellular Automation (CLA) and artificial neural networks (ANNs) were used for classification of rainy and no-rainy days. For classification of data, a comparison between CAL and ANNs indicated that accuracy of CLA with R2=0. 796 and RMSE=0. 407 is better than ANNs (R2=0. 556 and RMSE=0. 431). In order to predict daily rainfall, three methods were employed including ANNs with two learning functions of LM and BFGS and hybrid of ANNs-CLA. Although the ANN with learning function of LM predicts the rainfall with high accuracy (R2=0. 839 and RMSE= 0. 222), a hybrid of ANNs-CLA provides better results with higher accuracy (R2 = 0. 88 and RMSE= 0. 202)
Year: 2017