论文部分内容阅读
通过SVM分类机、朴素贝叶斯分类器和决策树算法,利用Matlab和WEKA软件以及粗糙集理论分析并验证了7个非线性因素对回采巷道围岩稳定性的影响。成功实现了对回采巷道围岩稳定性基于3种不同算法的训练和预测。从详细精度、混淆矩阵和节点错误率这3个方面分别比较了3种算法对回采巷道围岩稳定性分类预测的适用性。结果表明决策树算法不适用于对回采巷道围岩稳定性进行分类预测,SVM优于朴素贝叶斯分类器。
Using SVM classifier, Naive Bayesian classifier and decision tree algorithm, the influence of seven nonlinear factors on the stability of surrounding rock in mining gateway was analyzed and verified by Matlab and WEKA software and rough set theory. Successfully realized the training and prediction of surrounding rock stability of mining gateway based on three different algorithms. The applicability of the three algorithms to the classification and prediction of the surrounding rock stability of the mining gateway is compared respectively from the detailed precision, the confusion matrix and the node error rate. The results show that the decision tree algorithm is not suitable for the classification and prediction of surrounding rock stability in mining gateway. SVM is superior to Naive Bayes classifier.