Research on the Identification of Consumer Live Shopping Evaluation Features and Sentiment Classification
Constructing word vectors through Word2vec pre-training model,obtaining deep level feature information of context using Bi-GRU model,optimizing hidden layer weight information using multi head self attention mechanism,outputting features and emotion classification results using fully connected layer,integrating multi head self attention mechanism with Bi-GRU neural network model,exploring methods for recognizing and emotion classification of consumer live shopping evaluation features,Reveal the emotional connection between consumers and streamers,as well as the consumer's live shopping experience.
万方数据
