Detecting Depressive Symptoms on Social Media: A Comprehensive Review of Methodologies and Strategies for Suicide Prevention

Depressive disorder is one of the most underestimated and far-spread diseases of modern times, it can be misdiagnosed and remains undiagnosed for a good part of the sufferer’s life it often is seen as a precursor and one of the main symptoms for patients with severe disorders like that of borderline personality disorder, bipolar disorder which often leads to suicides taking more than 7,00,000 adult lives per year. Traits of depression can be easily judged by the personality traits which in turn can be accessed by how a person behaves and acts over multiple social media platforms, social media in this sense can be treated as a boon in terms of detecting people with sufferings. The following paper reviews and compiles various methodologies to segregate and classify people with depressive symptoms based on how they act on social media. Starting from basic thresholding techniques using dictionary-based methods right up till the complex models like LSTMs. It can be easily seen that quite extensive work has already been done in accessing people based on their social media profiles and researchers are still in process to develop more refined models. The aim of the following paper is to review techniques to check for profiles which are prone to suicides so that timely help can be provided.