Sentiment Analysis of Twitter for Detection of Depression using Machine Learning Algorithms

Abstract

Due to more internet users in the modern world, social media sites like Facebook, Instagram, and twitter have changed our world forever. Millions of people regularly share their opinions on social media that are related to different aspects of their lives. Over the years, stress, anxiety and depression are the most common psychological health issues that affect the people’s mind worldwide. This research study is focused on the detection of depression on twitter using various machine learning algorithms like SVM, Random forest, and Logistic regression. To develop the model, first tweets are downloaded from twitter and then perform the sentiment analysis on preprocessed tweets. After creating the final dataset, it is divided into 80:20 ratios to detect number of depressed tweets and non-depressed tweets using machine learning algorithms. After that, the performance of machine learning algorithms is analyzed and compared using different evaluation matrices like accuracy, precision, recall and F1 score. The results showed that, the highest prediction accuracy of 88% is achieved using Random Forest algorithm.