Predicting Stress Level in Online Education Due to Coronavirus Pandemic

Abstract

Education is one of the basic humans needs it is referred to as a country's backbone. Nowadays, Smart education technology has made the entire world smarter by the day. In just a few months, The Novel coronavirus (covid-19) has changed the world. As a result, the entire world is no longer relevant. In the context of Bangladesh, continuing one's studies while dealing with the pandemic is a new challenge. The role of education in our daily life is immense and in this pandemic time, there has been an unprecedented change in the field of education. The online education system has made education a novelty when the whole world has come to a standstill. This paper aims to find out the types of issues that a student may encounter in an online class during the covid-19 situation in Bangladesh. In this covid19 situation in Bangladesh, students continue their activities online and a student needs to analyze the types of issues that a student may encounter in an online class. In this paper, we used a machine-learning algorithm to and Data have been analyzed by basic questionnaires. To predict this problem, the accuracy was 74 percent given by Random forest by selecting all feature and 66 percent given by Xgb Boost by selecting Limited Feature. The motive of physical class and online class are almost similar to each other and it is very difficult to compare each class value. We used seven algorithms to run our model and the algorithm is KNN, Naive bias, decision tree, random forest, Xgb boost, SVM, and neural network. We have made use of Random forest classification algorithms, correlation heatmap, and univariant selection to select necessary data and to remove the unnecessary data that are not related to predicting students’ stress levels in online classes that have been discussed in a separate section of our paper.