Comparative Analysis of Clustering Algorithms and Moodle Plugin for Creation of Student Heterogeneous Groups in Online University Courses

Abstract

Online learning environments such as e-learning platforms are often used to encourage collaborative activities amongst students. In this context, group work is often used to improve the learning outcomes. Group formation is often performed randomly since university courses can be composed of a large number of students. While random formation saves time and resources, the student heterogeneity in terms of learning capabilities is not guaranteed. Although advanced e-learning platforms such as Moodle are widely used, they lack plugins that allow the automatic formation of heterogeneous groups of students. This work proposes a novel intelligent plugin for Moodle that allows the creation of heterogeneous groups by using Machine Learning. This intelligent application can be used in order to improve the students’ performance in collaborative activities. Our machine learning approach first uses clustering algorithms on Moodle data to identify homogeneous groups that are composed of students having similar behavior. Heterogeneous groups are then created by combining students selected from different homogeneous groups. To this end, a novel algorithm and the corresponding software, which allow the creation of heterogeneous groups, have been developed. We have implemented our approach by realizing a Moodle plugin where teachers can create heterogeneous groups.

Plant diversity and community analysis of Sele-Nono forest, Southwest Ethiopia: implication for conservation planning

“…5 ); and the optimum cluster number is quickly judged by inspecting the elbow (first breakpoint or bend) of the plotted curve with the assumption that adding another cluster doesn’t improve much better the total WSS (Woldu 2017 ). Researchers on the area suggest this elbow method as a fairly clear method, if not a naive solution as it is based on intra-cluster variance (Chebaeva et al 2020 ; Kongphunphin and Srivanit 2021 ; Nalli et al 2021 ).

Fig.

…”

Artificial Intelligence to improve learning outcomes through online collaborative activities

“…For group formation, an intelligent software, based on Clustering techniques using K-means algorithm applied to Moodle log data, creates firstly clusters of students with the similar characteristics and then heterogeneous groups distributing students of the same clusters into different groups (Nalli et al 2021). Moodle log data extracted by the e-learning platform allows for the calculation of various aspects of the student learning process, as well as the identification of students' behaviour using clustering techniques.…”

Formação de equipes heterogêneas: suporte ao docente para aplicação da Aprendizagem Baseada em Equipes

“…O tutor ou professor é responsável por analisar os clusters e grupos de estudantes gerados. Nalli et al, (2021) visaram à formação de grupos heterogêneos. Para isso, o trabalho primeiramente, realiza o agrupamento dos estudantes em clusters, considerando a similaridade do comportamento de cada estudante (instância) na interação com o AVA.…”

“…Nesse contexto, cada estudante é associado a uma classe ou categoria (perfil) conforme o cluster que o contém. Cada cluster recebe uma atribuição conforme suas características a exemplo de estudos de categorizações realizadas em [Oliveira et al 2022;Nalli et al, 2021;Pereira et al, 2021;Ramos et al,2020;Moubayed et al ,2020;Macedo et al ,2020;Pereira et al 2018]. Considerando a PAA ABE, tem-se os clusters, após a categorização dos clusters e sua atribuição a cada instância de estudante, a abordagem prossegue para a geração de grupos heterogêneos.…”