Suicidal ideation and mental disorder detection with attentive relation networks

Mental health is a critical issue in modern society, and mental disorders could sometimes turn to suicidal ideation without effective treatment. Early detection of from social content provides potential way for intervention. However, classifying other challenging as they share similar patterns language usage sentimental polarity. This paper enhances text representation with lexicon-based sentiment scores latent topics proposes using relation networks detect related risk indicators. The module further equipped the attention mechanism prioritize more relational features. Through experiments on three real-world datasets, our model outperforms most its counterparts.