采矿与安全工程学报2011，Vol.28Issue(3)：395-400,6.
深部硬岩岩爆评判的加权马氏距离判别法
A Weighted Mahalanobis Distance Discriminant Analysis for Predicting Rock-Burst in Deep Hard Rocks
摘要
Abstract
Distance discriminant analysis method, as one of the classification methods to estimate multi-index performance parameters, is now widely used for predicting and classifying rock-burst. However, the traditional Mahalanobis distance discriminant method redards the importance of all parameters as the same. So it exaggerates the role of some those parameters which have only small changes. According to the weighted Mahalanobis distance approach, the importance of each parameter is distinguished through the concrete application. Firstly, by means of analyzing the principal components of the parameters of rock-burst, this method calculates a new group of parameters and their corresponding percentage contributions, which are formed into the weighted matrix as a diagonal matrix of the percentage contributions. Then, by setting data to standardization, the weighted Mahalanobis distance can be calculated to classify rock-burst to be applied to practical project. Through contrasting the Mahalanobis distance discriminant method, the proposed method has a high discriminant ability and more accurate prediction results. Finally, rock-burst of hard rock in the Maluping and Dongguashan mines were predicted and evaluated under this model, whose results are almost consistent with the in-situ test results.关键词
岩爆/分类/加权马氏距离方法/预测Key words
rock-burst/ classification/ weighted Mahalanobis distance method/ prediction分类
矿业与冶金引用本文复制引用
王晋,李夕兵,杨金林..深部硬岩岩爆评判的加权马氏距离判别法[J].采矿与安全工程学报,2011,28(3):395-400,6.基金项目
国家重点基础研究发展计划(2010CB732004) （2010CB732004）
国家自然科学基金重点项目(50934006) （50934006）
湖南省研究生科研创新项目(CX2009B050) （CX2009B050）
