Relationship Between Student Engagement and Performance in E-Learning Environment Using Association Rules

Abstract

The field of e-learning has emerged as a topic of interest in academia due to the increased ease of accessing the Internet using using smart-phones and wireless devices. One of the challenges facing e-learning platforms is how to keep students motivated and engaged. Moreover, it is also crucial to identify the students that might need help in order to make sure their academic performance doesn't suffer. To that end, this paper tries to investigate the relationship between student engagement and their academic performance. Apriori association rules algorithm is used to derive a set of rules that relate student engagement to academic performance. Experimental results' analysis done using confidence and lift metrics show that a positive correlation exists between students' engagement level and their academic performance in a blended e-learning environment. In particular, it is shown that higher engagement often leads to better academic performance. This cements the previous work that linked engagement and academic performance in traditional classrooms.

Feature Evaluation of Emerging E-Learning Systems Using Machine Learning: An Extensive Survey

“…Second, they use a clustering algorithm to classify the extracted learner lines according to a specific learning style. In other approaches [87], the authors use the Felter-Silverman model as the LSM and FCM as the clustering mechanism, where fuzzy C means classifies the learner sequence with 96.89% accuracy and K-means classifies it with 80.12% accuracy, and thus the FCM clustering algorithm outperforms K-means [86] Clustering using K Means perform better in classifying student's groups based on student engagement level in classes [120].…”

“…The role of machine learning assistants has been analyzed and obtained in different studies. Supervised and semisupervised learning has strengthened research studies, and few studies have been conducted on supervised and semisupervised approaches based on the knowledge of researchers, revealing that the empowerment of learning methods similar to regression, regularization methods, instance-based methods, association rule learning, decision tree learning, deep learning, Bayesian approaches, clustering methods, kernel methods, association [120] E-Learning feature evaluations achieve good accuracy rates of up to 91% for large datasets through the advent of artificial neural networks and deep learning neural networks. Important e-learning feature measurement accuracy rates are summarized to compare and contrast the applicability of ML evaluations to fine tune the e-learning frameworks: In the analysis of machine learning algorithms available for improving the e-learning accuracy, a support vector machine outperforms other frequently used ML techniques such as kmeans, FCM clustering, and perceptron ANN model with a 97.15% accuracy rate in comparing the prediction of student knowledge levels and optimizing other parameters in elearning models.…”

Black-Box vs. White-Box: Understanding Their Advantages and Weaknesses From a Practical Point of View

“…Rule-based algorithms have widely been applied to different practical contexts, such as software defect prediction [123], inferring causal gene regulatory networks [124], evaluating the efficiency of currency portfolios [125], ranking of text documents [126], relationship between student engagement and performance in E-learning environment [127], assessing web sites quality [128], and exploring shipping accident contributory factors [129]; among others.…”

Uma Análise de Algoritmos de Clusterização para Descoberta de Perfis de Engajamento

“…A maioria dos estudos relacionados ao nível de engajamento tem-se pautado pela medição da frequência ou interação dos alunos. No estudo de [Moubayed et al 2018] utilizaram variáveis como notas do quiz, número de logins e nota do curso para medir o quão engajado está um aluno. No estudo de [Blanchettet, J.…”