A hybrid depressive mood analysis model to detect blogger depression tendency from web posts

Abstract

In recent years, reports of suicide have continuously increased due to people suffering from tremendous pressure or depression. According to World Health Organization (WHO), globally, 5% of adults suffer from mental disorders. This kind of mental disorder is difficult to diagnose and detect. In our previous works, we proposed a Negative Emotion Evaluation (NEE) model and an Event-Driven Depression Tendency Warning (EDDTW) model to detect depressive moods in advance. In this work, we combine the previous models to propose a Hybrid Depressive Mood Analysis (HDMA) model to predict depression from web posts. The experimental results show that our proposed hybrid depressive mood analysis model obtains over 70% precision.