comparison of artificial neural network (ann) and linear regression (lr) models to predict of forest and rangelands fires

natural fire inflicting irreparable damage to rangelands and forest areas is cause changes in landscape ecology. the purpose of this research is comparison of artificial neural network (ann) and line regression (lr) models to predict of forest and rangelands fires to this end, the data consist fire burned area and fire were used weather data over a period of 7 years (2006-2012(.the result indicates that the artificial neural network with implementation 9 and repeats 900 was obtained best performance network. the results of this study indicated the ability of neural networks in predicting the occurrence of fire and as well as the neural network model can to predict 86 percent change in forest and grassland fire using climatic parameters. also, the results of the spearman correlation test showed that the maximum temperature (p =0.006, r =0.896), minimum relative humidity (p =0.003, r =0.896) number of sunshine hours (p =0.010, r =0.876) levels correlated with fire area positively. the results of the linear regression between the burned and climatic factors were shown the coefficient of determination was 0.57. artificial neural network to predict fire in forests was more fast and reliable method than linear regression.