Abstract:To solve the over-fitting problem of Hopfield neural network,Hopfield neural network model with factor analysis was proposed based on factor analysis combined with Hopfield neural network model.Taking Dongliao River for an example,the model determined seven water quality evaluation factors using factor analysis method,created a 5×7 Hopfield neural network to evaluate water quality comprehensively,and compared the results from single Hopfield neural network and traditional Nemero Index method.Results showed that the factor analysis Hopfield neural network is much better than single Hopfield neural network.It not only makes up for the defect that factor analysis does not achieve the classification of water quality in practical applications,but also effectively reduces the extent of over-fitting of Hopfield neural network.The evaluation results are more reasonable and the model provides a new approach to comprehensive water quality evaluation with some excellent prospects.