节点文献
基于流形学习理论的冲击地压微震前兆辨识方法
Identification Method for Seismic Precursor of Rock Burst Based on Manifold Learning
【摘要】 针对微震信号的非线性、复杂性等特点,提出一种基于等距特征映射(Isometric feature mapping)和支持向量机(Support vector machine)相结合的微震前兆辨识方法。利用ISOMAP把数据从高维空间投影到低维空间而不改变数据内在属性的特点,对高维的微震信号进行降维,以发现嵌入在高维数据空间中的低维流形结构,作为冲击矿压灾变的最优敏感特征,采用Gauss核函数支持向量机作为分类器进行监测预警。实验结果表明,该方法可有效实现冲击地压微震前兆辨识。
【Abstract】 Aiming at the characteristic of microseismic signal by nonlinear and complexity,a method of Identification of seismic precursor based on Isometric feature map and SVM(Support vector machine)combination is proposed.For the advantages of ISOMAP that it can maintain the inner feature while projecting the high dimension data onto low dimension data space.To find embedded in high-dimensional data spaces in the structure of low-dimensional manifolds,which produced by ISOMAP from high dimension space of microseismic signal,and consider them as feature vectors,adopting gauss kernel function SVM classifier to monitoring and early warning.Experimental results indicate that this method can effectively achieve the identification of seismic precursor of rock burst.
【Key words】 Rock Burst; Microseismic Monitoring; Manifold Learning; Isometric Feature mapping; Support Vector Machine;
- 【文献出处】 软件导刊 ,Software Guide , 编辑部邮箱 ,2015年06期
- 【分类号】TP391.4
- 【网络出版时间】2015-06-23 16:46
- 【下载频次】71