General Information

This article introduces a recurrent CNN based framework for the classification of arbitrary length text in natural sentence. In our model, we present a complete CNN design with recurrent structure to capture the contextual information as far as possible when learning sentences, which allows arbitrary-length sentences and more flexibility to analyze complete sentences compared with traditional CNN based neural networks. In addition, our model greatly reduces the number of layers in the architecture and requires fewer training parameters, which leads to less memory consumption, and it can reach O(logn)O(log⁡n) time complexity. As a result, this model can achieve enhancement in training accuracy. Moreover, the design and implementation can be easily deployed in the current text classification systems.