ANN approach for modeling and prediction of water quality in Sichuan Kaschin-Beck disease districts

This paper establishes the nonlinear relationship predication model of artificial neural network (ANN) between the drinking water chemistry indicators and the morbidity of Kaschin-Beck disease in Sichuan districts by BP (Back-Propagation) arithmetic. Firstly, the input for the network are determined by the result of correlation analyses which analyses the water chemistry indicators and the morbidity of Kaschin-Beck disease, the output for the network is determined by the morbidity of Kaschin-Beck disease which have been obtained through to Zamtang county in Sichuan Province on-the-spot investigation. Then the hidden units for the network using the trial-and-error method, finally, the writer builds an ANN model for the simulation and prediction of water quality in Sichuan Kaschin-Beck disease districts, the structure of which model is 9:10:1. At last, the Model of artificial neural network calculates using the MATLAB. It achieved the high simulation precision and the desired results of training, then carries on the confirmation with the establishment model to the new sample, and it has obtained the high forecast precision. The results show that the model can establish the good relationship between Kaschin-Beck disease and the drinking water. The analyses made with this model show that the precision of both the simulation and prediction is high, so this model can be applied to forecast the morbidity of Kaschin-Beck disease. The model has great applied and popularized value.