摘要
                
                    提出一种改进的支持向量机模型,对电能质量扰动进行分类。支持向量机(SVM)在对大规模样本集的训练和分类时,需要占用大量内存,时耗过高,运算速度缓慢。针对这种情况提出一种改进的SVM模型:将原始训练样本集应用粗糙集理论(RS)去除冗余信息,然后在SVM中引入概率分布函数,用一个小规模的样本集训练得到一个初始的分类器,用这个初始分类器对大规模训练集进行修剪,修剪后得到一个规模很小的约减集,再用这个约减集进行训练得到最终的分类器。实验表明:这种改进的SVM模型有效降低了训练样本集的规模,提高了分类能力。
                
                A new support vector machine(SVM) model is proposed to classify power quality disturbances. The improved SVM method improves the speed of classification when SVM treats the large training set. Firstly, using rough set(RS) theory to eliminate redundant information of the large initial training set. Secondly, utilizing a probabilities function in SVM, training an initial classifier with a small training set and pruning the large training set with the initial classifier to obtain a small reduction set. Then, training with the reduction set, final classifier is obtained. Experiments show that this method effectively reduces the training set and improves the classification ability.
    
    
    
    
                出处
                
                    《电力系统保护与控制》
                        
                                EI
                                CSCD
                                北大核心
                        
                    
                        2010年第3期15-19,共5页
                    
                
                    Power System Protection and Control
     
    
    
    
                作者简介
俞晓冬(1974-),女,副教授,研究方向为人工智能在电力系统中的应用等;E-mail:xiaodongyu2001@163.com
周栾爱(1973-),女,研究方向为电力系统及其自动化等。