Enhancing Mental Health Awareness through Twitter Analysis: A Comparative Study of Machine Learning and Hybrid Deep Learning Techniques

Abstract

This study explores the utilization of social media data, specifically tweets and comments, for gaining insights into individuals' mental health conditions. The objective is to enhance mental health awareness and enable early detection and intervention. Twitter data is collected using depression-related keywords, and two models are employed: a Random Forest model with TF-IDF and a hybrid CNN-LSTM model incorporating word2vec. The performance of the CNN-LSTM model surpasses that of the Random Forest model, achieving an accuracy rate of 89.4%. Furthermore, a user interface is developed to analyze users' Twitter profiles based on their tweets, allowing for potential intervention through automated reply messages. By harnessing social media data and advanced machine learning techniques, this research contributes to improving mental health awareness and timely addressing of mental health concerns.