Student's Behavior in Virtual Learning Environment- A Case Study of Pakistan during Pandemic Situation of COVID-19

Abstract

Today one of the most challenging tasks is how to connect students with their education. During the Covid-19 era, the physical education system is not suitable for students. Most of the educational institutes start a new education system in Virtual Learning Environment. Suddenly changed the education system, the students towards learning environment is changed. The analysis of students through different machine learning and statistical techniques. The effectiveness of Virtual Learning is measured via the performance of the students. This research reviews different techniques for assessing the performance i.e activity-based, assessments of students, and enrollment-based. The analysis of students' behavior is a long-established task in the area of ML because in the past analyze the student's behavior in the statistical method. To compare, evaluate, and develop analyze the student's behavior in VLE,  we need a standard and high-quality benchmark corpus. But unfortunately, numerous studies are based on the web-based corpus and measure the performance in VLE. The main focus of this study is to analyze the student's behavior in VLE by using original data or collecting original reviews of students. Our total corpus consists of 2031 reviews. After some applying pre-processing technique final corpus consists of 1934 reviews. We applied seven machine learning algorithms to evaluate the student's behavior. To evaluate the performance of the students in VLE standard evaluation measures are used. After extensive experimentation, evaluation results show that stylometry word-based features produced the highest results of the first experiment NB Kernel (Accuray = 86.61, F1-measure = 89.91), and in the second experiment highest accuracy was achieved by DT (Accuray = 87.45, F1-measure = 92.13) on proposed corpus on reviews total 1934-Corpus.