Leveraging Neural Networks for Fine Grained Twitter Sentiment Analysis in the Context of Depression During Covid

The COVID-19 pandemic has significantly affected mental health globally, leading to a noticeable rise in anxiety, depression, and other psychological challenges. This study aims to explore the correlation between the pandemic and depressive symptoms expressed on social media, specifically Twitter. By applying fine-grained sentiment analysis using neural networks with a hybrid approach to word embedding, we analyzed tweets to detect depressive sentiments. The proposed model achieved an accuracy of 98.130/0, outperforming existing methods due to its ability to capture both contextual and sentiment nuances. This research provides insights into how digital platforms reflect societal mental health trends during crises and emphasizes the importance of real-time sentiment analysis for public health monitoring. Our findings indicate that the percentage of depressive sentiments in COVID19related tweets was significantly higher, reinforcing the link between global crises and deteriorating mental health.