Sentiment Text Spatio-Temporal Feature Analysis based on CharCNN and BiLSTM

This study employs social media data associated with Typhoon 'Lekima' as a case study to conduct text sentiment analysis by integrating text classification and sentiment analysis models. The goal is to investigate the patterns of user emotion distribution during the typhoon period. To initiate the process, a character-level Convolutional Neural Network (CharCNN) is utilized to effectively filter out irrelevant text data from Weibo content. Following this, the text content is represented using the word2vec model. Finally, a Bidirectional Long Short-Term Memory (Bi-LSTM) model is employed for sentiment analysis, enabling the extraction of emotional features from text sentences.