PCA-SVR在煤层瓦斯含量预测中的应用

针对煤层瓦斯含量与其影响因素之间存在着复杂的非线性关系,建立了基于主成分分析和支持向量回归机的煤层瓦斯含量预测模型。该模型有效地解决了小样本、非线性预测的问题,并发挥了主成分分析法消除输入变量间相关性的优点,减少了输入变量个数,提高了预测精度和收敛速度。通过实证分析,该模型的预测精度高,能够直接用于煤矿现场预测煤层瓦斯含量。

Application of PCA-SVR on gas content predicting in coal seam

In view of existing complicated nonlinear relation between gas content in coal seam and its influence fac- tors, a prediction model of gas content was constructed based on principal component analysis and support vector regression machine. The model can effectively solve the problems of small sample and nonlinear prediction; and makes use of principal components analysis to eliminate correlation between input variables, which reduces numbers of input variables to improve prediction precision and convergence rate. Through the empirical analysis, the predic- tion precision of this model was higher, which can be directly applied to predicting gas content in coal seam on the spot.