Research Of Text Sentiment Analysis Based On Attention Mechanism And Deep Learning

With the development of the internet industry,the promotion of e-commerce reviews and social platforms has accumulated a large amount of text data.Extracting the ideas and emotions contained in these information which can help people make some decisions and promotion.Text sentiment analysis has become a hot research topic in the field of natural language processing.Traditional text sentiment analysis model achieve good results based on the feature engineering of manual annotation by combining with grammar rules.With the development of the deep learning model in the processing of text sequences,the attention mechanism is introduced under the premise of feature engineering without manual annotation,the key parts of the text information are paid more attention,which has excellent performance in the task of text sentiment analysis.This paper mainly focuses on text sentiment polarity classification and different aspect text sentiment classification,and constructs the forward-reverse sequence AT-LSTM model and the deeper attention LSTM with aspect embedding(AE-DATT-LSTM)to handle text sentiment analysis tasks.In order to study the text emotional polarity classification,the AT-LSTM model is constructed to pay attention to the key information in the text based on the model of LSTM network,and the pre-trained Glove word vector is used.The reverse sequence is analyzed by the attention mechanism,then the feature fusion and classification processing are performed,then the paper put forward the AT-LSTM model of the forward and reverse sequences.The experimental results on the SemEval-2017 Task4 dataset show that the method improves the accuracy of text sentiment polarity classification in the text sentiment polarity classification task.Aiming at the emotional polarity classification of different themes,this paper put forward a new LSTM model,which combining aspect information and deeper attention.Through the bidirectional LSTM with shared weights,it trained the aspect embedding and the text embedding to get the aspect feature and text feature to carry on the feature fusion,and after the deeper attention mechanism processing,it obtained the classification result of the corresponding aspect by the classifier.The experimental results of the SemEval-2014 Task4 and SemEval-2017 Task4 datasets show that this method has further improved the accuracy and stability of the attention-based sentiment analysis model in the aspect sentiment analysis.The introduction of aspect features and deeper attention mechanisms is of great significance to the task of sentiment analysis based on aspect,which provides method support for public opinion analysis,text reasoning.