采空区危险性的支持向量机识别

针对采空区危险性影响因素与其危险性等级之间存在着复杂非线性关系的特点,笔者提出采用支持向量机最优分类理论来识别采空区的危险性等级。研究选取岩体结构、地质构造、岩石抗压强度、弹性模量、采空区形状、矿体倾角、高跨比、空区体积等8个参数作为主要影响因素,根据支持向量机理论,提出了1-V-1的采空区分类算法,并在Matlab中编程,建立了分类预测的SVM模型。以某矿山的实测采空区为例,利用该模型进行了识别,并与BP神经网络预测结果作对比。实例研究表明,采用该方法的分类结果比神经网络更准确,与采空区调查结果一致性好,用支持向量机理论进行采空区危险性评价是可行的。

Recognition of goaf risk based on support vector machines method

Complicated non-linear relationship is existed between the grade of goaf risk and its influence factors. In order to classify the grade of goaf risk, support vector machines classification method is studied. In this paper, rock structures, geological structure, rock compressive strength, elasticity modulus, shape of goaf, ore body angle, engage in collapse, volume of goaf are treated as main influencing factors. A 1-V-1 classification algorithm is proposed based on SVM, as well as SVM models are built for classification in Matlab. Taking one underground mine as an example, the paper uses the SVM models to recognize goaf risk and compares the result with that of BP neural network prediction. It shows that the classification of SVM method is more accurate than that of neural network method, and there is a high agreement with the survey results. SVM is a feasible method on goaf risk evaluation.