Multi-label Classification of Short Texts with Label Correlated Recurrent Neural Networks

Zhiyuan Peng, Behnoush Abdollahi, Min Xie, Yi Fang

Abstract: Short texts are commonly seen nowadays on the Internet in various forms such as tweets, queries, comments, status updates, snippets of search results, and reviews from social platforms. Accurate categorization of these short texts is critical for enhancing information services as it provides the foundation for better search and recommendation. In many real-world applications, a short text is often associated with multiple categories. Due to the sparsity of context information, traditional multi-label classification methods do not perform well on short texts. In this paper, we propose a novel Label Correlated Recurrent Neural Network (LC-RNN) for multi-label classification of short texts by exploiting correlations between categories. We utilize a tree structure to represent the relationships among labels and consequently an efficient max-product algorithm can be developed for exact inference of label prediction. We conduct experiments on four testbeds and the results demonstrate the effectiveness of the proposed model.