Intelligent identification and fusion prediction of coal,rock and gas dynamic disaster risk
Coal,rock and gas dynamic disaster poses a great threat to mine safety production.Due to many inducing factors,its inherent disaster mechanism is difficult to be effectively explored.In order to realize intelligent identification and prediction of coal,rock and gas dynamic disaster risk,an intelligent identification and fusion prediction model based on CNN was established.In the model,Box-plot and MI methods are used for data cleaning,and GRA method is used to establish an indicator system of coal,rock and gas dynamic disaster including 10 risk factors.After dimensionally reducing the data by PCA method,the data are input into the CNN model for fusion and prediction.The comparison and analysis with ANN,BP,RF and SVM models show that the intelligent identification and fusion prediction model of coal,rock and gas dynamic disaster risk based on CNN has higher accuracy,and the convergence speed of this model is faster,which verifies that this model has more reliable engineering value in practical engineering.
coal,rock and gas dynamic disasterconvolutional neural networkprediction modelrisk identificationdeep learning
万方数据
维普
