An Efficient Machine Learning Classification with Feature Selection Techniques for Depression Detection from Social Media

The emergence of Facebook, Instagram, and Twitter among other social media sites has profoundly altered our daily routines. With more people online than ever before, it's easy to establish a unique persona in cyberspace. The positive aspects of social media are undeniable, but the negative aspects of its use are also rather evident. More time spent on social media is associated with a higher risk of getting depression, according to recent research. Depressive symptoms include a persistently low mood and a loss of interest in previously enjoyed activities. Distinct mental health issues may have far-reaching effects, and severe depression, sometimes called major depression, is no exception. This research looks at two distinct machine learning classifiers that can tell whether a person is sad based on various demographic and psychological factors. Further, feature selection techniques such as recursive feature elimination with linear regression have select the most important characteristics from the dataset. RFC and GBC were utilized to improve the precision with which depression could be predicted. This technique is a machine learning (ML)-based prediction strategy for diagnosing depression and other mental diseases early. In addition, numerous models have had their F1-score, sensitivity, accuracy, specificity, and AUC calculated to see which is best for identifying episodes of depression. Higher accuracy (96%) was shown in diagnosing of depression using the ML algorithm.