Applying Deep Learning in Depression Detection

Abstract

According to the World Health Organization, one in twenty people in the world have suffered from depression and emotional distress in the previous twelve months. How to manage and provide appropriate treatment to people suffering from depression and emotional distress is a highly pressing issue. However, many people with depression and emotional distress are not sufficiently recognized and treated and do not actively seek help. It is therefore highly desirable to devise a method to effectively and proactively identify these people. Following the design science approach, we propose DK-LSTM, a novel design based on deep learning to identify people with depression and emotional distress. Based on Long Short-Term Memory (LSTM), a type of deep learning networks, our model incorporates both general knowledge and domain knowledge in the learning process through word embedding and parallel LSTM units.