摘要
文章通过SVM分类机、朴素贝叶斯分类器和决策树算法,利用WEKA等软件以及粗糙集等理论分析并验证了瓦斯放散速度、瓦斯压力、煤的普氏系数、地质破坏程度、开采深度等非线性因素对煤与瓦斯突出的影响。成功实现了对煤与瓦斯突出基于三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三个方面分别比较了三种算法对煤与瓦斯突出分类预测的适用性。结果表明:三种算法对煤与瓦斯突出进行分类预测是较成功的,其中,决策树算法对煤与瓦斯突出进行分类预测的效果最优,其次为朴素贝叶斯分类器。
This article analyzes and verifies the effects of nonlinear factors,including gas diffusion velocity,gas pressure,sturdiness coefficient of coal,geologic destroy degree and mining depth,on coal and gas outburst by means of the SVM sorting machine,naive bayesian classifier and decision tree algorithm,and uses WEKA software and the rough set theory. It has successfully realized the training of and prediction for coal and gas outburst based on three different algorithms. Furthermore,it also has compared the applicability of the three algorithms for the classification and prediction of coal and gas outburst from three aspects,including detailed accuracy,confusion matrix and the error rate of node. The results show that it is successful to classify and predict the coal and gas outburst by using the three algorithms among which the naive bayesian classifier has the best effect and that of decision tree takes the second place.
出处
《煤》
2016年第4期23-25,共3页
Coal