Explainable Detection of Depression in Social Media Contents Using Natural Language Processing

Abstract

This paper an explainable deep learning approach using Long Short-Term Memory (LSTM) networks to detect depression from social media posts. The model classifies text into depression or control categories by capturing linguistic patterns and sequential dependencies. An attention mechanism is integrated to enhance interpretability, highlighting key features influencing predictions. Evaluated on a public mental health dataset, the model shows high accuracy and transparency, offering a scalable solution for early depression detection and supporting timely mental health interventions.