Analysis on the Accuracy of Different Classification Prediction Models in the Field of the Impact of Online Classes on Students During the COVID-19

Abstract

Some students are pleased, while others are sad, as a result of the introduction of online education. Different attitudes, levels of physical health, family economic conditions, and so on have resulted from online classes. Consequently, it is necessary to determine what has caused various students to have different states and whether or not parents of school-aged children took the appropriate response measures. In this paper, the author preprocesses relevant datasets from Kaggle, then uses naive Bayesian, random forest, K-neighbor, SVC, logistic regression, and neural networks to classify and predict the dataset, analyses their accuracy and confusion matrices to determine the best classification and prediction model, and conducts importance analysis based on the best classification and prediction model. Therefore, it is possible to determine the significance of each parameter and to make suggestions based on the significance of various parameters. This can be used to determine whether there is room for development and errors in the next plans implemented in the era of online classes, and in the future, it can be used to reduce the health problems of students caused by the next pandemic in the era of online classes.