estimation of nitrate in hamedan-bahar plain groundwater using artificial neural network and the effect of data resolution on prediction accuracy

information on nitrate in groundwater resources requires periodic measurements are accurate. despite the measure in some areas due to sensitive social and health community are not reported. therefore, be informed of the status of each area of water quality, modeling is essential. the purpose of this study was the application of artificial neural network method for estimating nitrate and compared with measured and estimated effectiveness of nitrate from the number and type of input data to neural network models. data from 53 groundwater wells hamedan - bahar plain, two groups of costly information and low cost, during the years 2003 to 2008 were collected. in costly information, of the 13 independent variables were used as chemical input neural network and in low-cost group of seven and eight variables separately for modeling nitrate was used. comparison of three structures indicates the high ability of neural network models in predicting the nitrate concentration. comparison of the average error from all three neural network models with t test and z statistics showed significant differences between the model results, there isn't. therefore, the input data in neural network group is justified. model input parameters include the depth of the static characteristics of geomorphology, deep wells, geographical and qualitative information of temperature, ph, ec of water samples was measured that was predicted, with nitrate concentrations of more than % 80 confidence that shows model performance is good in the aquifer of hamedan– bahar.