EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

3708. Application of Predictive Analytics in Assessing Student Performance: A Case Study in a Computer Supported Collaborative Learning Environment

Invited abstract in session MD-45: Emerging Trends in Decision Analysis, stream Decision Support Systems.

Monday, 14:30-16:00
Room: 30 (building: 324)

Authors (first author is the speaker)

1. Yannis Psaromiligkos
Business Administration, University of West Attica
2. Athanasios Spyridakos
Department of Business Administration, University of West Attica
3. Natasa Themeli
University of Athens
4. Theodoros Anagnostopoulos
University of West Attica
5. Christos Kytagias
University of West Attica
6. Dimitris Papakyriakopoulos
Business administration, University of West Attica

Abstract

In contemporary collaborative learning environments, characterized by intricate activities and numerous stakeholder interactions, there's a pressing need to gather insights from diverse sources to attain a comprehensive understanding of learning dynamics and students' behaviours. While Learning Management Systems serve as the cornerstone of these environments, their data analysis capabilities are often limited to basic metrics such as visit frequency, completion rates, and score statistics. Consequently, there's a high demand for specialized tools to support instructional designers. Decision support in the form of Learning Analytics, rooted in the domain of business intelligence, has emerged as a forefront area of research in technology-enhanced learning. In this paper, we present the application of predictive analytics within a multi-criteria dataset structure to categorize students based on their performance in collaborative activities. Our approach was deployed in a real-world scenario involving two cohorts of an undergraduate course conducted entirely online due to the pandemic. The first cohort comprised 337 students organized into 81 groups, while the second cohort consisted of 341 students divided into 76 groups. We applied various machine learning algorithms to determine the algorithm yielding optimal performance.

Keywords

Status: accepted


Back to the list of papers