Narrative Review on Depression Detection in Online Social Media

This narrative review examines the research papers published between 2017 and 2022 that investigate depression detection through the analysis of social media. The review comprises a thorough search of databases of five prominent digital libraries with the goal of identifying high-quality research contributions pertaining to the identification of depression. After that, ten pertinent papers were chosen and carefully examined. In the papers that we have examined, the most common way that data is collected is by using the APIs of social media platforms, with Twitter being the most used platform for this purpose. Term Frequency-Inverse Document Frequency technique was most frequently used to extract speech features. The studies discuss the relationship between mental health issues and public health risks in the setting of depression detection. Additionally, they demonstrate how the use of deep learning and ML algorithms can consistently aid in recognizing and reducing the difficulties presented by psychological disorders such as depression, hence improving public health management.