基于不同判别准则的硬岩矿柱状态识别模型  

Recognition model of hard rock pillars state based on different discriminant criterions

作　　者：赵国彦[1] 刘建 周健[1] ZHAO Guoyan;LIU Jian;ZHOU Jian(School of Resources and Safety Engineering,Central South University,Changsha 410083,China)

机构地区：[1]中南大学资源与安全工程学院,湖南长沙410083

出　　处：《中南大学学报（自然科学版）》2018年第11期2813-2820,共8页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(51374244)~~

摘　　要：基于162个硬岩矿柱样本,构建不同判别准则下矿柱状态识别的Fisher判别分析(FDA)模型、距离判别分析(DDA)模型和Bayes判别分析(BDA)模型,进而与多元逻辑回归(MLR)、极限学习机(ELM)、最小二乘支持向量机(LS-SVM)、支持向量机(SVM)、高斯过程分类(GPC)、分类回归树(CART)、神经网络(ANN)共7种常用的统计学习方法进行比较,同时探讨主成分分析(PCA)方法提高识别准确率的可行性,并对矿柱状态影响因子进行敏感性分析。研究结果表明:这10种统计学习方法中,GPC的准确率最高,FDA的准确率次之,然后是MLR,CART的准确率最低;对于3种判别分析方法,FDA的准确率最高,DDA与BDA的准确率几乎相当;增加判别指标,DDA和BDA的判别准确率显著降低,其他方法对判别指标增减不敏感;对某些方法,原始数据经PCA处理后不能提高其判别准确率;矿柱状态对矿柱应力最敏感,其次是矿岩单轴抗压强度,其对矿柱宽高比的敏感性较低。Based on 162 hard rock pillar cases,three discriminant models for pillar stability determination including Fisher discriminant analysis(FDA)model,distance discriminant analysis(DDA)model and Bayes discriminant analysis(BDA)model were constructed,and then they were compared with 7 statistical learning methods which were multiple logistic regression(MLR),extreme learning machines(ELM),the least squares support vector machines(LS-SVM),support vector machines(SVM),Gaussian process classification(GPC),classification and regression tree(CART)and artificial neural network(ANN).At the same time,the feasibility of increasing the recognition accuracy with principal component analysis(PCA)was discussed and sensitivity analyze of each influence factor were carried out.The results show that the best three methods are GPC,FDA and MLR,CART has the lowest accuracy,and the accuracy of DDA and BDA is nearly equal.The accuracy of DDA and BDA decreases with the increase of the number of influence factors,but accuracies of other methods are not sensitive to the change of the factor numbers.For some methods,the prediction accuracy does not increase as expected when the raw data are processed by PCA.Pillar state is most sensitive to the pillar stress,followed by the uniaxial compressive strength of the rock,and the sensitivity to the ratio of pillar width to its height is low.

关 键 词：硬岩矿柱 状态识别 判别分析 主成分分析 敏感性分析 

分 类 号：TD853[矿业工程—金属矿开采]