Sentimental prediction model of personality based on CNN-LSTM in a social media environment

The paper first analyzes the correlation between text sentiment values and personality traits, proves that text sentiment can have a good support effect on user personality prediction, then on this basis, a method based on CNN-LSTM is proposed, which can be used to deeply analyze the sentiment analysis capability of the model, hoping to improve the precision of sentiment classification and lay a solid foundation for the next experiment. This experiment proves that the CNN-LSTM constructed in this paper can better predict the emotional tendency of the short text of microblog, has good generalization ability, and has higher precision than other methods.