关键词: 地质环境, 支持向量机（SVM）, 环境预测

Abstract: The Yushenfu mining district, located in transition region between the hill ravine area of Loess Plateau in the north of Shaanxi province and Maowusu sand land, has many characteristics: the embedding depth of coal seam is shallow, the mining seam is thick and the thickness of overlying base rock is rather thin, in addition, the surface in this zone was covered by rather unconsolidated formation. Because of the drought and lack of rain all of the year and the sparse vegetation, it is a typically ecological vulnerability. Consequently, the large-scale coal mining easily result in some more serious problems of the geological environment. The research analyzes the geological mining factors and study the influence on geological environment exerted by coal mining. Adopting the theory and methods of the support vector machine (SVM), we built an assessment of integrated geological environmental quality and a nonlinear prediction model. By evaluating and forecasting the evolution result of nonlinear interaction of geological environment factors on coal mining, we got predicting results from five different comprehensive geological environment divisions and the mining change divisions. The method has a more scientific and more accurate effect, closer to the reality in evaluating the nonlinear interaction that comes from the factors of complex geological environment and predicting the evolution of comprehensive geological environment.

Key words: geological environment, support vector machine (SVM), environmental forecast