COVID EmotionNet: A Machine Learning Approach to Unraveling Pandemic Sentiments

Currently, COVID-19 is recognized as one of history's most significant health epidemics. Maintaining mental health and physical health are burning questions when a crisis occurs like COVID-19 disease. During pandemic situations, people express their mental and physical conditions through their comments, posts, tweets, and pictures on social media because of the advancement of the internet. Researchers have already devised different models to extract human mental and physical health information. Within this Investigation, we devise a framework leveraging individuals' posts on Twitter during the COVID-19 crisis. We engage five m odels i n t he r ealm of machine learning: Random Forest, Naive Bayes, Decision Tree, Logistic Regression, and Support Vector Machine (SVM), and we also deployed three ensemble methods Bagging, Boosting, and Stacking. We also used two-word embedding techniques TFIDF and Word2vec to normalize the dataset. The proposed approach can detect human emotion regarding COVID-19 disease. The model is trained using the Kaggle Coronavirus tweets NLP Text Classification dataset. The finding sh ows that the stacking ensemble method with TFIDF vectorization offers the best result (Accuracy-88.25%, Precision-89%, Recall-87%, and F1-Score-88.2%) to detect human emotion about COVID-19.