矿井突水水源的SVM识别方法

发布时间：2018-05-29 19:38

本文选题：矿井突水 + 水源识别　；参考：《辽宁工程技术大学学报(自然科学版)》2017年01期

【摘要】：为解决在有限监测信息下,矿井突水水源识别中原始指标众多,识别矩阵呈现出的高维、稀疏性问题,采用相关性理论和支持向量机(SVM),提出了矿井突水水源识别方法.计算训练样本相关性及单指标误判个数.依据误判数对各指标进行重要性排序.构建指标组合并约简出测试正确率最高的指标组合,再利用构建的基于相关性理论的矿井突水水源SVM识别模型,实现高维小样本评价.结果表明:Ca~(2+),Mg~(2+),K~++Na~+彼此间相关性较高;K~++Na~+和CL~-,HCO_3~-和SO_4~(2-)相关性均较低;Ca~(2+)较重要,SO_4~(2-)和CL-较不重要;约简评价体系的冗余指标,可节约监测成本,且能够在有限监测信息下,保证识别正确率.
[Abstract]:In order to solve the problem of high dimension and sparsity in mine water inrush water source identification under limited monitoring information, a method of mine water inrush water source identification is proposed by using correlation theory and support vector machine (SVM). The correlation of training sample and the number of single index misjudgment are calculated. The importance of each index is ranked according to the number of misjudgments. The index combination is constructed and the index combination with the highest test accuracy is reduced. Then the SVM identification model of mine water inrush source based on correlation theory is constructed to realize the evaluation of high dimension and small sample. The results show that the correlation between K ~ Na ~ + and K ~ Na ~ + is higher than that between K ~ (Na ~ and CL-HCO3- and so _ 4 ~). The correlation of K ~ Na ~ and CL-HCO3- and so _ 4 ~ (2) are lower than that of Cash2) and CL-), and the redundant indexes of the evaluation system can save monitoring cost and can be used in limited monitoring information, the results show that the correlation between K ~ Na ~ (2) and K ~ (Na ~ (2) is higher than that between K ~ (2) and K ~ (2). Ensure correct recognition rate.
【作者单位】：辽宁工程技术大学工商管理学院;
【基金】：国家自然科学基金项目(51274110)
【分类号】：TD745

【相似文献】