Applying radial basis function(RBF) neural network to predict the sediment deposited from check dam

Three indicators (R, I
30
, P), and all four indicators (R, I
30
, P, I) of erosive rainfall in Jia Zhaichuan small watershed of Song county are chosen respectively as the input vector to predict sedimentation volume with the two neural network of RBF and BP, and fit with the actual values. The results testify the fitting and predicted effects of RBF neural network are all better than BP network, as well as the indexes (R, I
30
, P) are the main factors causing soil erosion.