Multi-view semi-supervised classification overview

With the continuous exploration of data collection technology, the multi-view data, obtained from different ways or levels for the same object, and the multi-view algorithm, used for modeling and solving problems from multiple perspectives, have attracted extensive attention. In addition, if the traditional supervised learning method is used in multiview learning, a great amount of labeled data are required to ensure the accuracy of the training model, but it is difficult to obtain the labeled samples. Thus, combining semisupervised learning with multi-view method can effectively solve the difficulties of labelling. In this paper, we will describe the related multi-view semi-supervised learning algorithms and development status in the aspect of classification, and then introduce the latest real world applications. Finally, we will summarize and point out the possible research directions in the future.