煤巷围岩分类的Bayes判别分析法

选用影响煤巷围岩分类的主要6项指标作为Bayes判别分析模型的判别因子,以15组煤巷围岩实测数据作为学习样本进行训练,通过分析计算,建立了相应的线性判别函数,并利用回代估计方法对训练后的判别模型进行检验,其回判估计的误判率为0。利用该模型对3组煤巷围岩实测数据情况进行识别,其结果与实际情况完全一致,并与神经网络和支持向量机的预测结果进行对比。结果表明：Bayes判别分析法判别能力强,交叉确认估计的误判率低,是解决煤巷围岩分类的一种有效方法。

Method of Bayes discriminant analysis for classifying mine lane surrounding rocks

The Bayes linear discrimination model was established through training 15 practical measured data,in which six factors were chosen as variables;and the ratio of mistake distinguish,was considerably zero after the discrimination model was trained.3 practical measured data were used to verify this model,the prediction accuracy was acceptable,and its contrast with the neural network and support vector machines.The results show that the Bayes discriminant analysis method has excellent discrimination potency and a low cross validated misjudgment ratio,is a good method using in the research on classifying mine lane surrounding rocks.