Semi-supervised multi-label Boosting algorithm
一种半监督的多标签Boosting分类算法

For multi-label classification problem without enough labeled data, this paper proposed a new semi-supervised Boosting algorithm. It provided a semi-supervised general multi-label Boosting framework by using functional gradient descent method. It also used the conditional entropy as a regularization term on unlabeled data in classification model. Experimental result shows that the performance of the new semi-supervised Boosting algorithm can be improved by increasing unlabeled data; it also has a better result than traditional supervised Boosting algorithm by different measures.