Chinese text sentiment analysis based on improved Convolutional Neural Networks

Abstract

Convolutional neural networks(CNN) has achieved fairly good results in the field of computer vision. In recent years, with the rapid development of deep learning, more and more researchers tried to apply CNN to the field of Natural Language Processing(NLP). This paper uses CNN model to analyze the sentiment of Chinese text, and improves the structure of basic CNN, adjusts different parameters to carry out multiple sets of parallel experiments at the same time. The result is that dual-channel network model is more accurate than the single channel model, and the accuracy are 93.4% and 92.7% respectively where the word vector comes from character-level and word-level. Which shows it is feasible to apply CNN to the field of Chinese sentiment analysis. For Chinese text, the effect that word vector based on the character-level is much better than that based on word-level. And the dual-channel CNN model provides a new idea for the further exploration in the field of NLP.