An English Text Classification Method Based on TextCNN and SVM

Text classification is an important problem in natural language processing. The text representation is a matrix with high sparseness and high latitude. However, traditional text classification requires manual feature engineering with subjective interference factors. Therefore, an English text classification method combining a convolutional neural network and support vector machine is proposed in this paper. This method uses convolutional neural networks’ advantages to extract features, avoiding feature engineering in traditional classification methods. Meanwhile, the proposed way fully plays the generalization performance of support vector machines. The Internet Movie Database (IMDB) movie review data were applied to evaluate the performance of the proposed model, and the experimental results showed that the proposed approach achieves an accuracy rate of 0.9062, which exceeds the other three methods.