基于证据理论的不平衡数据半监督分类方法

摘要

针对标记样本少的不平衡数据,提出一种基于证据理论和biased-SVM的不平衡数据半监督分类方法.首先采用随机子空间方法得到不同的视图;然后在各个视图上利用初始的标记样本集训练biased-SVM模型,并将其应用于未标记样本集,从而得到未标记样本的类概率输出;最后引入证据理论进行信息融合来提高标注的稳定性.通过在一些公共数据集上的实验结果表明,与其他方法相比,在不同的标记样本率下,所提方法均具有较高的数据集整体的G-mean值和小类的F-value值,并具有较高的稳定性.%This paper proposed a semi-supervised classification method based on evidence theory and biased-SVM for imbalance data sets which had a number of unlabeled samples.First,the method used the stochastic subspace method to get different views.Second,it trained biased-SVM model using the initial labelled samples on each view,then the trained model was applied to unlabled samples to get probability outputs.At last,it adopted evidence theory to improve the stability of unlabeled samples signatures.Experimental results on some public data sets show that compared with other methods,the proposed approach can more effectively and stably utilize the unlabeled examples to improve the value of G-mean and minority class F-value under the different rate of labelled sample.