A Sentiment Analysis Method Based on BLSTM and CNN Fusion
Liedong Yin, Ye Xia, Junping Yao · Journal of Physics Conference Series · 2018
Sentiment analysis is a research hotspot in natural language processing in recent years. This paper proposes a sentiment analysis method that integrates LSTM and CNN in view of the fact that most existing sentiment analysis methods mix semantics and emotions. This method divides the text into semantic space and emotional space, and uses the attention model to train in two dimensions, and fuses the feature representations of the two to construct the final representation of the text. The results show that the model has achieved good results in multiple emotional review data.