论文部分内容阅读
针对SVM训练学习过程中难以获得大量带有类标注样本的问题,提出一种基于距离比值不确定性抽样的主动SVM增量训练算法(DRB-ASVM),并将其应用于SVM增量训练.实验结果表明,在保证不影响分类精度的情况下,应用主动学习策略的SVM选择的标记样本数量大大低于随机选择的标记样本数量,从而降低了标记的工作量或代价,并且提高了训练速度.
In order to solve the problem that it is difficult to obtain a large number of labeled samples in SVM training and learning process, an active SVM incremental training algorithm (DRB-ASVM) based on distance ratio uncertainty sampling is proposed and applied to incremental training of SVM. Experimental results show that the number of labeled samples in SVM using active learning strategy is significantly lower than that of randomly selected labeled samples without affecting the accuracy of classification, which reduces the workload or cost of marking and increases the training speed .