SMOTE和Biased-SVM相结合的不平衡数据分类方法

Imbalance Data Set Classification Using SMOTE and Biased-SVM

针对不平衡数据集的分类问题,本文利用支持向量机推广能力强的优良特性,提出了SMOTE(Synthetic Minority Over-sampling Technique,SMOTE)和Biased-SVM(Biased Support Vector Machine,Biased-SVM)相结合的方法。该方法首先对原始数据使用Biased-SVM方法,然后对求出的支持向量使用SMOTE向上采样方法进行采样,最后再使用Biased-SVM方法进行分类。实验结果表明,本文采用的SMOTE和Biased-SVM相结合的方法可提高不平衡数据集分类精度。

In view of the classification of the imbalance data set,this paper gives the method using SMOTE （Synthetic Minority Over-sampling Technique, SMOTE） and Biased-SVM （Biased Support Vector Machine, Biased-SVM）. Firstly, data set is taken into account using Biased-SVM algorithm. Secondly, the sampling to support vector is carried on using the method of SMOTE upward sampling. Finally, the classification is carried on using the method of Biased-SVM. The experimental result indicates that the method given in this paper improves the precision of imbalance data set classification.