Non point source pollution in the reservoir of Hedi conforms to simulation study

Based on the investigation crane reservoir and made pollutants based GPC new method for predicting the opposite. Jiuzhou River from 2000--2010 years of monitoring data as training and test samples. Non-point source pollution load factors and the influence of nonlinear mapping relationship between the structure by GPC learning and training samples. Monitoring data Jiuzhou river since 2000-2010 experiments demonstrate the impact of the above methods. Compared with other machine learning methods, it suggests, GP method is more feasible, effective and simple.