Decoding depression: Analyzing social network insights for depression severity assessment with transformers and explainable AI

Tasnim Ahmed, Shahriar Ivan, Ahnaf Munir, Sabbir Ahmed

Abstract: Highlights•Proposed a Transformer-based ensemble architecture to classify depression severity.•Improved input quality and relevance using diverse preprocessing techniques.•Analysis from local and global perspectives to explain decision-making strategy.•Compared the performance of zero-shot LLM (ChatGPT) with the proposed architecture.