Depression Predicting Model For Social Media User’s Emotions Using Deep Learning Methods

Abstract

There is a growing interest in leveraging social media platforms for the early identification of depression among the youth due to a rising occurrence. The study investigates the feasibility of utilizing Twitter data specifically emojis are included along with text content to predict degrees of depression among its users. The study examined Twitter data and used machine learning and deep learning to predict depression. Previous research on emotion detection mainly focused on text alone, but the contemporary landscape of communication integrates both text and emojis. The significant rise in usage of emojis in underlines their crucial role in emotional expression, necessitating their inclusion in analytical processes. The study shown the impact of combining text with emojis to determine the polarity of expressed sentiments. The study employed Convolutional Neural Networks (CNN) for sentiment analysis, with promising findings in reliably recognizing negative feelings and the potential for high performance to sentiment categorization. The study created a model capable of not only recognizing depression levels within a dataset but also identifying people who have higher-than-average levels of depression. This model acts as an early warning sign for mental health support, enabling mental health practitioners and social media platforms to offer timely assistance and interventions.