Depression Detection in Twitter Tweets Using Machine Learning Classifiers

Abstract

Users are increasingly using social media to share their thoughts and feelings. One of the primary causes of higher suicide rates is 'depression.' A user’s social media status recently revealed information on their mental health, as well as their current circumstances and actions. Various studies have been conducted to determine how social media may be used to assess a user’s mental health, and the results have been impressive. This is accomplished by the examination of expressed views, images, attitudes, language style, and other activities. The uploaded tweets from Twitter users will be used in this study to detect depressions. Five machine learning algorithms were employed to discriminate between depressed and non-depressed tweets. Support Vector Machine, Decision Tree, Random Forest, CNN, and Naive Bayes are all examples of machine learning algorithms. Finally, with 85.00 percent accuracy, 81.25 percent precision, 90.00 percent recall, and 82.90 percent f1-factor, Decision Tree Classifier exceeds all of the examined and evaluated approaches. This research can serve as a foundation for intelligent system developers working in the field of detecting user mental states.