Text-based Depression Detection on Social Media Posts: A Systematic Literature Review

Due to the huge increase of awareness of mental health well-being, the detection of mental illness itself is starting to become a huge concern. Many psychiatrists found difficulties in identifying the existence of mental illness in a patient because of the complicated nature of each mental disorder, thus making it hard to give the appropriate treatment to the patient before it’s too late. However, due to the integration of social media into people’s daily life, this create an environment that may provide additional information regarding the mental disorder a patient bear. This study has been undertaken as a Systematic Literature Review (SLR), which is defined as a process of identifying, assessing, and interpreting the available resources to provide answers for a set of research questions. Analysis is made to answer questions regarding text-based mental illness detection based on the social media activity of people with mental disorders, and reveals that it indeed is possible to do early detection of depression on social due to the existence of a particular characteristics in the way these subjects use their social media. This SLR found that from the small amount of research using text-based approach, most studies use deep learning models such as RNN on the early detection of depression cases due to the limitation of data availability. However, this study will look to find method that may prove to be more effective.