Detecting Depression on Social Platforms Using Machine Learning

Abstract

Depression is the most common disease in the world. However, a sizable portion of depression-related illnesses go untreated because they are not properly diagnosed. Social media provides a platform to assess individuals’ mental conditions. Therefore, social media offers an innovative method to identify possible sad individuals. Previous research has shown that it is possible to analyze social media posts made by people with major depressive disorder in order to determine whether they are currently experiencing or are going to experience depression. People are increasingly using microblogging platforms like Twitter and Reddit to broadcast their thoughts and activities. The research community has presented a number of methods to quickly detect those who are depressed on social media. Performance can be enhanced in comparison to the current research’s low accuracy and limited use of phrases, sentences, and blog posts. The goal of this study is to explore whether machine learning can be used to accurately identify depressive symptoms in social media users by analyzing their posts—especially when those messages don't directly contain phrases associated with depression. Our study seeks to highlight a machine learning (ML) classifier that effectively detects depression in users of social media. Five machine learning classifiers—SVM, NB, KNN, DT, and RF—were used in conjunction with preprocessing techniques to train all of the classifiers on Reddit and Twitter in order to accomplish the goal. Then, using Reddit and Twitter information, we analyzed the models and determined which one is more accurate for diagnosing depression. The results show that the SVM classifier, with 76% on Reddit and 78% on Twitter, is the best ML classifier that produces the greatest results across multiple social networking sites.