Recurrent Attention Capsule Network for Text Classification

The main advantage of convolutional neural networks is the extraction of key local features with strong expression capabilities, and usually requires a large amount of training data to achieve better performance. However, the convolutional neural network ignores the hierarchical relationship between the local features and the spatial structure, which may result in a lower accuracy of text classification. For this problem, a text classification method based on Recurrent Attention Capsule Network is proposed. The proposed method uses the capsule network to obtain the spatial position relationship of local features, and uses the recurrent structure to obtain the context information of the words, and uses the attention mechanism to further obtain more important key features for classification. Compared with CNN network and RNN network in text classification, the experimental results show that the experimental accuracy of the proposed method is better than six benchmark algorithms.