The web-based behavior of online learning: An evaluation of different countries during the COVID-19 pandemic

Abstract

COVID-19, a pandemic that the world has not seen in decades, has caused several new obstacles for student learning and education throughout the globe. As a consequence of the worldwide surge of COVID-19 instances, several schools and institutions in almost every region of the globe have closed in 2020 or switched to online or remote learning, which will have a variety of repercussions for student learning. This has led to educators and students spending more time online than ever before, with both groups researching, learning, and familiarizing themselves with information, resources, tools, and frameworks to adapt to online or remote learning. Data mining and analysis are being done to analyze such online activity. For the construction of this dataset, the web-based data in the form of search interests connected to online learning, gathered through Google searches, was mined using Google Trends. Currently, the dataset comprises web-based data related to online learning for the 20 nations that COVID-19 negatively touched at the time of its construction. This project aims to create and evaluate time-series forecasting models of a country's end-of-term performance, explore how the pandemic has influenced the migrations of people throughout the globe, and estimate the nations' future online learning needs. Regression techniques such as linear regression, multilayer regression, and SMO regression are utilized. This is done by looking at previous data, identifying the trends, and creating short-term or long-term projections. The data demonstrate that the approach of SMO regression causes fewer errors with improved accuracy compared to others.