利用支持向量机识别松辽盆地火山岩岩性

利用支持向量机(SVM)方法,选取个性特征元素,建立火山岩岩性成分的识别方法,来区分玄武质、安山质、粗面质、英安质、流纹质火山岩岩性.通过对松辽盆地内部的火山岩样本进行学习和预测,火山岩大类平均识别率达到95%以上,表明支持向量机在火山岩岩性成分识别方面取得了良好效果.

RECOGNITION OF THE LITHOLOGY OF VOLCANIC ROCKS IN SONGLIAO BASIN BY SUPPORT VECTOR MACHINE

Using the method of support vector machine（SVM）, with selection of characteristic elements, an identification method for the lithology of volcanic rocks is established to distinguish the basaltic, andesitic, trachytic, dacitic and rhyolitic volcanic rocks. By learning and prediction of the volcanic rock samples from the Songliao Basin, the average recognition rate for volcanic rocks reaches to 95% and more, showing that the SVM obtain a good result in the identification of volcanic rock component.