【摘要】 为了有效而准确地分析智能电网中输电线路故障和寻找诱发故障的主要因素,研发了基于二叉树的核密度逻辑回归多分类模型,以解决输电线路不对称故障分析的问题。该模型根据Nadaraya-Watson密度估计将训练数据映射到了特定的特征空间,根据二叉树结构特点将多个DLR模型组合成一个具有多分类能力的二叉树,并加以优化。实验结果表明,基于MCDLR的分类结果在准确率上和分类时间上明显优于已有的传统的多分类算法。更多还原

【Abstract】 This paper proposes a novel MCDLRBT model(Multi-Classification of Density logistic regression based on binary tree,for short,MCDLRBT)in order to solve the problem of asymmetric fault analysis of transmission lines.The MCDLRBT model firstly maps the training data to a specific feature space based on the Nadaraya-Watson density estimation.Then,according to the structure characteristics of the two forked tree,MCDLRBT combines the multiple DLR(Density Estimation Logistic Regression,DLR)model into a two fork tree with multiple classification abilities and optimizes it in the end.The experimental results show that the classification results based on MCDLRBT are superior to the existing traditional multiple classification algorithms in the respects of accuracy and classification time.更多还原