The application of bayesian approach to discrimination of mine water-inrush source

Discriminating the water-inrush source correctly and quickly is an important guarantee for mine safe production.The paper selected many items of water parameters for each aquifer and applied Bayes method to establish the model of rapidly identifying the water-bursting source for different types-water coal mine.Using SPSS, the example of Huainan Guqiao mine was compared with the fuzzy comprehensive evaluation and neural network model.The results show that Bayes multi-class linear discriminant model has a higher discrimination accuracy than the fuzzy comprehensive evaluation and has the same accuracy as neural network model.However, Bayes linear discriminant model is superior to neural network model because of its simple calculation and stable structure.