The coal and gas outburst prediction model research based on SVM

被引：0

作者：

Chen, Shan-Le ^{[1
,2
]}

Liu, Jian ^{[2
]}

Chen, Liang ^{[3
]}

机构：

[1] College of Safety Science and Engineering, LiaoNing Technical University, LiaoNing, FuXin, China

[2] BiJie University College of Resources and safety engineering, GuiZhou, BiJie, China

[3] Chongqing Research Institute co. Ltd of China Coal Technology and Engineering Group Corporation, Chongqing, China

来源：

International Journal of Earth Sciences and Engineering | 2014年 / 7卷 / 02期

关键词：

Iterative methods - Coal - Forecasting - Accidents - Gas bursts - Gases;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

Coal and gas outburst accidents always result in heavy casualties and property losses, and it is difficult to forecast because the parameters in prediction model are complexity. As there are many samples from historical coal and gas outburst accidents, we considered 5 factors of coal and gas outburst in this paper; they are gas pressure, firmness coefficient, and initial velocity of gas radiation, damage types and mining depth. We forecast the coal and gas outburst by the model that based on SVM. The results show that the model based on SVM can provide effective and accurate prediction. Otherwise, we discussed the choice of kernel function and the obtaining of the model parameters, through iterative calculation respectively. The results suggest that the Square Sine kernel function is the best fit for coal and gas outburst prediction model, and the optimal model should include the natural support vectors, and the more the percentage of the boundary support vector, the more prediction accurate of the model. ©2014 CAFET-INNOVA TECHNICAL SOCIETY.

引用

页码：616 / 623