Emotions Prediction From The Text of Social Media Using NLP and Different ML Models

Abstract

This paper "Emotions Prediction From The Text of Social Media Using NLP and Different ML Models." describes the application of state-of-the-art natural language processing and deep learning methods to emotions conditions using social media. This work identifies the emotions and sentiments signaling the bad emotional conditions, such as anxiety, depression, and stress, using both English and Bengali textual data of social media posts. Advanced transformer models, including BigBird, ERNIE, and XLM-R, are able to process long sequences of text, even in multilingual sets. These models were fine- tuned using labeled data for supervised training, with sentimental and emotiona annotations that will help in increasing the accuracy of the classification. The results of these models showed that among them, ERNIE had the best performance in terms of accuracy and F1 score, having an accuracy of 95.87% and an F1 score of 93.08%. These findings from the research demonstrate the feasibility of using deep learning models for real-time monitoring of emotional condition through social media, which may offer a rather accessible and non-invasive way for early detection of mental health issues. The project shows how AI can contribute Emotions prediction From the Text of social media using NLP and Different ML Models.