Sentiment Recognition of Web Text Data Based on Multiple Attention and Long- and Short-Term Memory Networks

In this paper, in order to realize the accurate recognition of the sentiment of web text data, multi-head attention and long- and short-term memory networks are brought into the data sentiment recognition technique. The key information contributing to the entity is quickly filtered out, and the contextual global information is extracted from front to back and back to front, and the dependencies between neighboring words in a single sentence are obtained. Through the text data preprocessing received by the artificial neural network and through the network text data emotion feature extraction, calculate and analyze the real emotion of the network text data and its classification and modeling, and finally the recognition structure is also the real emotional intent of the network text data. The results show that the accuracy of this paper’s method for emotion recognition of network text data and the recognition rate of each class are both 0.99, and the test time takes 36[Formula: see text]s to accurately recognize a variety of emotions of network text. It is able to recognize the emotion of network text data in a fast and real-time manner and has great value for popularization and application.