基于灰色支持向量机的矿山安全事故预测研究

为了提高矿山安全事故预测的可靠性,在灰色预测模型GM(1,1)和支持向量机SVM的基础上,提出了矿山安全事故次数的灰色支持向量机G-SVM的组合预测模型.首先采用GM(1,1)进行数据趋势预测,然后对于残差序列采用支持向量机预测进行捕获,最后将两种模型的结果进行融合,得到组合预测结果.结果表明,组合模型比单一的GM(1,1)模型和SVM模型具有更高的预测精度.

Study on the prediction model of mine safety accidents based on the grey support vector machine

Due to the characteristics of more random nature and limited samples in mine safety accidents,the single forecast model is inadequate to achieve the ideal prediction accuracy. In order to improve the reliability of mine safety accident prediction,the combined forecast model based on the grey forecasting model( GM( 1,1)) and support vector machine( SVM) was proposed. At first,GM( 1,1) was used to obtain the possible data,and then with the support vector machine( SVM),the residual sequence was predicted. Finally,the comprehensive forecast results can be obtained by adding the prediction values of two established models together. The results show that the combined model has higher prediction accuracy compared with the prediction accuracy of single GM( 1,1) model and the SVM model. The result of the combination forecast can provide an effective method for mining safety accident prediction research.