针对电能质量监测系统的海量多特征数据信息,提出采用基于支持向量机的回归特征消去法进行特征选择,综合支持向量机对不同的电能质量特征集的分类正确率选取了最优特征集。以高速铁路电能质量数据为例,利用该方法对有无高铁负荷运行进行了分类研究。实验结果表明,所选出的特征集反映了高铁电能质量特点并具有很好的分类效果,证明了所提方法的可行性,为电能质量数据挖掘分类提供了一种思路和方法。 In view of the massive multi-characteristic data of power quality monitoring system, the regression feature elimination method based on support vector machine is proposed to select features. The comprehensive support vector machine (SVM) selects the optimal feature set for the classification accuracy of different power quality feature sets. Taking high-speed railway power quality data as an example, this method is used to classify whether there is high-speed railway load operation. The experimental results show that the selected feature set reflects the characteristics of high-speed rail power quality and has good classification results. It proves the feasibility of the proposed method and provides an idea and method for power quality data mining classification.