Affective Classification of Chinese Product Reviews by Combining Text Content and Labels
In order to solve the problem of insufficient utilization of language knowledge and label information in the emo-tion analysis methods based on deep learning,in this article we improve the emotion analysis method based on deep learning,and propose a Chinese text emotion analysis method that combines text features and label information.In our im-proved method we first preprocess the text,and then use syntax rules and deep learning methods to extract the feature infor-mation of the text respectively.In the part of syntax rules,we extract the more explicit information of emotion tendency in the text by constructing syntax rules,and combine it with the label description information to further highlight the emotion information.In the part of deep learning,we extract the high-level features of the text by combining the attention mechanism and gated recurrent neural network.Finally,we combine the features extracted from the above two parts and use the classifier to classify them to get the emotional orientation of the text.The experiment showed that the method proposed in this article was better than the traditional emotion analysis methods.
emotion analysisdeep learningsyntactic ruleslabel information
万方数据
维普
