基于不同文本表示协同训练的半监督文本分类算法

In the task of semi-supervised text classification, co-training, based on the differentiated feature sapce, has achieved good results with the use of supervised classifiers. However, the way to find the dual-view from content to meet the conditions of full redundancy and conditionl independence is the difficulty of text co-training. In this paper, two different feature spaces are constructed from different text representation models, which are based on different points and ways. As the double views of co-training, the particularity for scenes in the existed models are solved. On this basis, an improved co-training algorithm for unbalanced dataset is also presented. The experimental results show that the proposed co-training model is superior to semi-supervised text classification.