Twitter Data-Driven Depression Detection using Machine Learning

Abstract

These depressive symptoms are plaguing today's society, and the number of those afflicted is steadily rising. Unfortunately, not everyone who is struggling with depression is aware of it; in fact, some of them are able to recognize it. Conversely, many are using social media as a "diary" to document their emotional condition. Using machine learning algorithms, many types of study have been carried out to identify cases of depression in social media posts. Researchers can learn whether social media users are suffering from depression by analyzing the data that is publicly accessible. In order to distinguish between data that is depressed and data that is not, a machine learning algorithm can sort the data into the appropriate categories. The purpose of the proposed study is to identify user sadness using social media data. The Twitter data is then put into two distinct kinds of classifiers, which are Naïve Bayes and a hybrid model, NBTree. The findings will be evaluated based on the greatest accuracy value to select the best algorithm to diagnose depression. The findings demonstrate that both algorithms are equally effective, since they achieve the same degree of accuracy.