Detecting At-Risk Students With Early Interventions Using Machine Learning Techniques

Abstract

Massive Open Online Courses (MOOCs) have shown rapid development in recent years, allowing learners to access high-quality digital material. Because of facilitated learning and the flexibility of the teaching environment, the number of participants is rapidly growing. However, extensive research reports that the high attrition rate and low completion rate are major concerns. In this paper, the early identification of students who are at risk of withdrew and failure is provided. Therefore, two models are constructed namely at-risk student model and learning achievement model. The models have the potential to detect the students who are in danger of failing and withdrawal at the early stage of the online course. The result reveals that all classifiers gain good accuracy across both models, the highest performance yield by GBM with the value of 0.894, 0.952 for first, second model respectively, while RF yield the value of 0.866, in at-risk student framework achieved the lowest accuracy. The proposed frameworks can be used to assist instructors in delivering intensive intervention support to at-risk students. INDEX TERMS Machine learning, massive open online courses, receiver operator characteristics, area under curve.

Comparison of Logistic Regression and Generalized Linear Model for Identifying Accurate At – Risk Students

“…GLM is a statistical method that comes under supervised machine learning algorithms which assumes every number of observations has a distribution like polynomial, binomial, gamma, average (Al-Shabandar et al 2019). GLM gives a continuous output of dependent variables by evaluating independent variables and performs linearly.…”

“…(Geigle and Zhai 2017), Two-layer hidden Markov Model was used for predicting accuracy percentage of At-risk students on the basis of student's behavioural patterns and results demonstrated that Low motivational status, withdrawal of course at early stage and failure rate are the major concerns and got an accuracy of 80.93%. (Al-Shabandar et al 2019), At-risk students accuracy was predicted using machine learning algorithms which includes GLM, Gradient Boost machine algorithm, Random Forest. The results of their work demonstrated an accuracy of 91%, 90%, 91% respectively.…”

“…The results of their work demonstrated an accuracy of 91%, 90%, 91% respectively. The most cited article is proposed by the author (Al-Shabandar et al 2019) focused on GLM with an accuracy of 91%.…”

“…Now-a-days many colleges, Universities and schools have adapted the digital education systems as an alternative to the traditional education system. Everyone prefers the digital education system (Al-Shabandar et al 2019). Many research studies have proved that High failure rate and withdrawal rate were the major concerns of the digital education systems (J.…”

Predictor Analysis of Middle and Low Achieving Students in Cell Biology Online Course in Health Nutrition Undergraduate Program

“…However, potentially low achieving students are more di cult to detect in an online course, since the lecturers and students are not within the same room and the time for online face to face contact is limited and often shorter than in-class meetings. Therefore, early detection of low achieving students will provide more opportunity to overcome some of their challenges and problems in achievement [6,7].…”