A self-attention TCN-based model for suicidal ideation detection from social media posts

Seyedeh Leili Mirtaheri, Sergio Greco, Reza Shahbazian

Abstract: Highlights•Urgent need for early suicide ideation detection highlighted due to WHO reports.•Shared content on social networks enables AI-based early suicide ideation detection.•Proposed model distinguishes suicidal ideation among posts with high accuracy.•TCN, self-attention, and LSTM enhance detection capabilities in proposed model.•Proposed method results over 94% accuracy, recall, and F1-score.