Topic-Sentiment Analysis of Public Health Events Network Opinion Based on KBert and BiLSTM-ATT

This article proposes a topic-sentiment analysis model based on KBert (Bert+K-means) and BiLSTM-ATT, which studies the public opinion of public health events network from the perspective of text topic and sentiment evolution. The model combines the strengths of KBert, a novel method that integrates BERT with K-means clustering, and BiLSTM-ATT, a powerful architecture for sequence modeling with attention mechanisms. By analyzing both the evolving themes and sentiment expressed within online discussions, this study aims to provide valuable insights into public opinion regarding public health events on social media networks.