Estimation of Learners' Levels of Adaptability in Online Education Using Imbalanced Dataset

The COVID-19 epidemic has had a huge impact on education, causing a quick move to online learning environments. Students had to adjust to a virtual learning environment as a result of this move, which brought new obstacles for them. The study looks at how adaptable students are in online learning and how it affects their academic performance and general learning process. This study aims to assess students' levels of adaptation to online learning settings and to pinpoint the factors that either support or hinder students' ability to adapt to this new learning environment. According to preliminary data, with respect to personal circumstances, students' adaptation in online learning differs greatly. The dataset used for the study is students_adaptability_level_online_education.csv is taken from Kaggle Machine Learning repository. The minority class in the dataset is equalized using the Synthetic Minority Oversampling Technique (SMOTE), as it comprises an unbalanced set of values. In order to equalize the distribution of data, the values for the minority class are increased at random. Dataset has been employed to predict the degree of student adaptability to online education through a number of machine learning algorithms, including Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbours (KNN) and xgb.XGBClassifier(). The Random Forest classifier had the greatest accuracy (92%), compared to those that were employed. The demand for customized and adaptable learning experiences grows as online education continues to transform the traditional learning environment. This study addresses the idea of student adaptivity in the education with an emphasis on how it may significantly enhance learning results for students from a variety of backgrounds and learning styles.