Sentiment analysis of Chinese Weibo combining BERT model and Hawkes process

Abstract

In order to effectively monitor the changes in the user's emotional state and perform sentiment analysis on their Weibo, this paper proposes a Chinese Weibo sentiment analysis algorithm that combines the BERT (Bidirectional Encoder Representations from Transformers) model and the Hawkes process for sentiment prediction. The algorithm first uses the BERT model to calculate the sentiment score of the user's microblog, and then uses the Hawkes process function to predict the sentiment using the score obtained. Experimental analysis and comparison show that the algorithm has achieved good experimental results in sentiment analysis and prediction of Chinese Weibo.