Application of Logistic regression model based on rough set in recognition of mine water inrush pattern

The mine water bursting pattern recognition is a non normal, nonlinear and high dimensional data processing problem, but also a binary-class problem. The attribute reduction algorithm of rough set was used to reduce the dimension of the sample data, to establish Logistic regression model, and particle swarm algorithm was used to optimize model parameters. The recognition accuracy of the model was 90% for water inrush mode of the modeling samples and 100% for water inrush mode of the testing samples, the effect was better than that of the Logistic regression model without dimensionality reduction. The model overcomes the shortcomings of the linear regression analysis for the solution of the binary-class problem, provides a new method for pattern recognition of mine water inrush.