Forecasting model of mine construction project accidents based on SOM neural network

To improve the control capability of mine construction enterprises on their projects´ safety management, a forecasting model of mine construction project accidents based on SOM neural network is presented on the analysis of influencing factors using its characteristics of coping with non-linear problems. The correct prediction results indicates that the model can overcome the disadvantages of contaminated data samples with resort to its particular abilities of adaptability and learning for models, showing the strong robustness when the forecasting model is applied to accident prediction as long as the influencing factors is firmly grasped. It also can enhance the safety management performance of mine construction project.