Abstract: Classification is one of the key techniques of data mining. It requires a large number of training samples to oblain a favorable classifier, but it is resource-consuming to create label for each sample, it is even more so for multi-label samples. In order to reduce costs, it should find the most informative samples which can represent the classes. The classificrs which arc based on SVM, the larger margin, the classifier's accuracy will be poorer. hhis paper proposed an acfive learning method based on SVM's expect margin which relies on current classifier, select samples that can reduce classifier's margin fastest hhe experimental results show that the method based on expect margin outperforms than other active learning strategy based on decision value and posterior probability strategy.

Key words: Multi-label, Posterior probability, Expect margin, Active learning, SVM