Sentiment analysis of Sina microblog comments related to fertility issues based on GNN-LSTM model

This study utilized a GNN-LSTM model for sentiment analysis of microblog comments related to fertility issues. A dataset of approximately 7300 comments was collected through a Python crawler, cleaned, and segmented. The model was trained on the CCF Big Data and Computational Intelligence competition's training data and fine-tuned. TF-IDF text feature extraction was applied to transform the text data into feature vectors, achieving an F1 score of about 0.81 on the training set. For model testing, the test set's text data was used for sentiment analysis, and sentiment orientation vectors were visualized with a three-dimensional scatter plot. WordCloud technology generated word cloud maps to reveal keyword and sentiment word distribution in the data. The findings highlighted that discussions on fertility issues predominantly revolved around marriage and financial concerns, with most netizens expressing neutral or negative emotions. This suggests external factors like policy and social environment may contribute to declining fertility rates. The study acknowledges limitations in handling complex text, such as irony, and calls for further research and innovation to address these challenges.