<p>Online education is one of the educational solutions for students during the COVID-19 pandemic. Worldwide, most universities have shifted much of their learning frameworks to an online learning model to limit physical interaction between people and slow the spread of COVID-19. The effectiveness of online learning depends on multiple factors, including student and instructor self-efficacy, attitudes, and confidence in using the technology involved; the educational strategies employed; the ability to monitor and evaluate educational outcomes; and student motivation, among many others. Therefore, in the present work, we will be assessing the performance and student condition in the online mode of education by predicting the degree to which students are satisfied in online mode depending on factors like internet facility, interaction in online mode, etc. We will be evaluating the data, which is collected using a survey method with almost 1000 records with 22 attributes by applying various machine learning models (XGBoost, Random Forest, Catboost, Gradient Boosting) and a deep learning model named TabNet to predict the impact of online education among the students. To enhance the accuracy of the prediction model, we have ensembled the machine learning and deep learning models. The ensemble model gave an accuracy of 86.6%, with an F1 score of 0.863 on the dataset.</p>