Spatial Distribution of Water Quality Based on Generalized Regression Neural Network

The paper proposes a new model for analyzing spatial distribution of water quality. The model is based on statistics and generalized regression neural network. The model is based on statistic theories and easy to approximate a nonlinear function. The advantages of the model include simple net structure, few artificial parameters, and high accuracy calculation. The developed approach is successfully applied to the water quality assessment of the Taihu Lake of China P.R, which includes the dissolved oxygen concentration of water-quality index and the spatial distribution of the water-quality level. The results show that the model can generate a credible spatial distribution of the water quality.