Towards improving online education system in Bangladesh using machine learning approach

Abstract

In this rapid advancement of technology online education has come out as an alternative to traditional classroom based learning systems especially in the contexts like Bangladesh. While Bangladesh embraced online education during the COVID-19 pandemic there remain critical challenges to ensure its quality and equitable access. This research finds out the ways to improve the standard of online learning from Bangladesh's distinct point of view and proposes strategies for improvement considering the country's distinct socioeconomic and cultural background. Firstly it investigates the infrastructure issues that have a substantial impact on the success of online learning efforts such as internet connectivity and digital device access. Secondly the study examines the pedagogical strategies used in the delivery of online learning highlighting the importance of engaging and culturally appropriate teaching strategies. The study aims to determine tactics that maximize learning results and student engagement in the online environment by looking at best practices and empirical data. For the purpose of training and evaluating our dataset, we can use the Random Forest Classifier (RF), Decision Tree Classifier (DT), K-Nearest Neighbor (KNN), Logistic Regression (LR), Naïve Bayes, Support Vector Classifier (SVC), and XGBoost Algorithm (XGB). Highest accuracy gain Random forest 98% and lowest accuracy linear Logistic Multiclass,Naïve Bayes, 43% . Accuracy obtain in KNeighbors Classifier 82%, Accuracy obtain in SVC 70% , Accuracy obtain Linear SVC 44% Accuration obtain in Ensemble 85% .