Simulation of nitrate contamination in groundwater using artificial neural networks

摘要: In this study, performance of two artificial networks was evaluated to determine which one would have more efficiency in predicting nitrate contamination groundwater. The case study Babol is recognized as the most fertile regions Iran. Relevant factors including hydrogeology, soil nitrogen content, organic matter and carbon content were measured situ input data predict groundwater, then correlated by using Pearson formula. Next, back-propagation radial basis function neural applied one-by-one. best structure for model found be 4-5-1 Radial with a spread parameter equal 0.5 mean square error (MSE) 0.50 mg/l. Results showed no significant difference between proposed models. Both ANN models can reliably groundwater acceptable accuracy. However, marginally better compared 30 %.