中文摘要：

标准支持向量机（SVM）对不平衡数据集进行分类时,会出现不平衡现象;传统不平衡数据集分类方法只能对不平衡数据集分类,且在分类过程中存在人工因素的参与。提出一种平衡不平衡数据集统一分类方法——自调节分类面支持向量机（self-adjusting classification-plane SVM,SCSVM）,设计自适应的分类面调节方法,根据训练错分情况对分类面进行调整,控制正负类样本的错分率使其达到均衡,平衡或不平衡数据集都可采用相同的方法进行分类而不需预知数据集种类。实验表明该方法可对平衡或不平衡数据集进行有效的分类。

英文摘要：

The imbalance phenomenon occurs while imbalance data set is classified by normal SVM; traditional methods for imbalanced data set only deal with imbalance data set and there has artificial experience participating in classification. A new classification method based on self-adjusting classification-plane SVM is proposed which can process balance and imbalance data set in unified classification form, classification-plane of SVM is adaptively adjusted according to the training samples classification error, the classification error rates of positive and negative samples is controlled to reach equilibrium for positive and negative classes. The experiments show that the method is effective in the classification of balance or imbalance data set.