Abstract
Sentiment analysis leverages machine learning and natural language processing techniques to classify and interpret textual data, identifying sentiments as positive, negative, or neutral. This study explores sentiment analysis in the context of mental health, utilising two models: Logistic Regression and Bidirectional Encoder Representations from Transformers (BERT). The dataset comprises 52 680 unique statements associated with seven mental health statuses, including depression, anxiety and suicidal tendencies. Logistic Regression achieved an accuracy of 72%, while BERT, with its advanced deep learning architecture, demonstrated a significant improvement with an accuracy of 84%. BERT’s superior performance is attributed to its bidirectional contextual understanding and attention mechanisms, enhancing its ability to handle complex and nuanced textual information. This study highlights the efficacy of BERT over traditional models in analysing and classifying sentiments related to mental health, underscoring its potential for improving early detection and intervention.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call