A machine learning approach for student assessment in E-learning using Quinlan's C4.5, Naive Bayes and Random Forest algorithms

Student Assessment on e-learning platforms is a debated subject. The focal emphasis of this research study is to predict fair/transparent student evaluation using machine learning algorithms. A prediction on students' final grade showing whether the student will pass or fail would benefit the student/instructor and act as a guide for future recommendations/evaluations on performance. An in depth study on the assessment techniques for e-learning such as Markov Model, metacognitive perspectives has been conducted. A proposed model for fair/transparent student evaluation/performance has also been presented. Specific parameters have been defined that are then efficaciously tested by applying machine learning algorithms. In this study, classifiers such as Decision Trees-J48, Naive Bayes and Random Forest are used to progress the excellence of student data by initially eradicating noisy data, and consequently getting better prognostic accuracy. The scope of the paper has been set for undergraduate programs. The experimental results endow with set of guidelines to those students who have low grades. Performance testing has also been conducted for verification, accuracy and validity of results.