基于ν-SVR算法的岩爆预测分析

以预测地下工程岩爆发生为研究目的，在综合影响岩爆的关键因素的基础上，选取地下工程围岩最大切向应力、岩石单轴抗压、抗拉强度、弹性能量指数、围岩切向应力与围岩抗压强度比值、围岩抗压强度与其抗拉强度的比值作为岩爆预测的评判指标，建立了一种基于改进支持向量机算法( ν-SVR)的岩爆预测方法，并利用国内外45个岩石地下工程实例进行学习，对另外的16个实例进行了预测，取得了较好的效果，其预测精度明显优于灰色理论和常规SVR算法，与GA-BP神经网络算法相近.

Rockburst prediction analysis based on ν-SVR algorithm

In order to predict the rockburst occurrence of underground engineering,according to the collected data from actual underground rock project,selecting the wall rock's maximal tangential stress,rock's single axle ten- sile strength,rock's single axle pressive strength,elasticity energy index,the ratio of rock's single axle pressive strength and rock's single axle tensile strength,and the ratio of wall rock's maximal tangential stress and rock's single axle pressive strength as the judging indexes of rock burst,a method for rockburst predicting model based on v-SVR(support vector regression)was put forward.Applied the predicting model to predict rockburst of 16 un- derground rock engineering after learning with other 45 samples;the result is satisfactory.It is more accurate than a gray theory and classical SVR,and is resemble with GA -BP neural network algorithm.