摘要
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。
To the SVM sorting machine, naive bayesian classifier and decision tree algorithm, and uses WEKA, Matlab softwares and the rough set theory to analyze and verify the influences of seven nonlinear factors on the stability of mining roadway. The training and prediction of the stability of surrounding rock in mining roadway based on three different algorithms is triumphantly realized. From three aspects of detailed precision, confusion matrix and node failure rate to compare the applicability of three algorithms for the classification prediction of stability of surrounding rock in mining roadway respectively. The results have shown that the decision tree algorithm is not suitable for classification prediction of the stability of surrounding rock in mining roadway and the SVM is superior to the naive bayesian classifier.
出处
《煤炭技术》
CAS
北大核心
2015年第7期46-49,共4页
Coal Technology
基金
国家重点基础研究发展规划(973计划)(2012CB723104)
中国煤炭工业协会指导性计划项目(MTKJ2012-345)
关键词
回采巷道
围岩稳定性
SVM
朴素贝叶斯分类器
决策树
mining roadway
stability of surrounding rock
SVM
naive bayesian classifier
decision tree