论文部分内容阅读
针对电能质量(power quality,PQ)复合扰动识别中缺少特征选择与最优决策树自动构建方法的不足,提出采用分类回归树的PQ特征选择与最优决策树构建方法.首先,通过S变换提取64种PQ特征,构成原始特征集;然后,采用嵌入式特征选择方法,获取特征Gini重要度及排序,确定最优特征子集;最后,应用1-标准误差规则子树评估法,进行代价复杂度剪枝,获得最优分类树.实验证明,新方法能够根据训练集自动构建最优决策树,并实现最优特征选择;最优决策树可准确识别不同噪声环境下,含多种复合扰动的PQ信号,分类准确率高于概率神经网络和支持向量机方法,具有良好的鲁棒性与抗噪性.“,”The lack of feature selection and optimal decision tree automatic construction method in complex power quality disturbances identification, a novel feature selection and optimal decision tree construction method based on classification and regression tree (CART) was proposed. Firstly, the 64 features of power quality were extracted by S transform to construct the original feature set. Then, the optimal feature subset were selected by Gini importance and sorting using an embedded feature selection method. Finally, one standard error rule subtree evaluation methods were applied to cost complexity pruning. After pruning, the optimal classification tree was obtained. The experimental results show that the new method can automatically construct the optimal decision tree and achieve the optimal feature subset selection according to the training set. The optimal decision tree can accurately identify power quality signals with multiple kinds of complex disturbances in different noise environments. The classification accuracy is higher than probabilistic neural network (PNN) and support vector machine (SVM). The new method has good robustness and anti-noise performance.