L6-Article de recherche-2

Type de publication	Article de recherche
Livrable associé	L6
Titre de la publication	Process Mining Model to visualize and analyze the Learning Process
Date de publication	2021-12-17
Auteur	Maria Moreno , Ernesto Exposito, Mamadou Gueye
Lien vers la publication	https://aaee.net.au/wp-content/uploads/2021/11/REES AAEE 2021 paper 311.pdf

CONTEXT In online learning environments, the teacher provides students with a learning path to follow in order to acquire the expected competencies and skills. However, students' profiles are different as they can learn according to different learning paces or media content. Therefore, the actual learning path followed by each learner may vary from the initial path provided in the learning management system (LMS). This paper proposes an analysis of the learning paths followed by the students in order to identify and promote the most adapted learning processes in order to improve competencies and skills acquisition.

PURPOSE OR GOAL The learning traces left by students in their learning environment could be exploited in order to better understand and guide learning processes. Unfortunately, with large-scale education, the analysis of different learning paths can be a complex task to be manually carried out by teachers. For this reason, our objective is to propose an approach to model, visualize, analyze and recommend the most efficient learning process in order to improve students’ education experience and results.

APPROACH OR METHODOLOGY/METHODS The approach adopted is based on the learning traces left by the students following their interactions with the Learning Management System (LMS). After collecting, processing, and storing these learning traces, Process Mining technologies are used to analyze the data through an exploration of the learning process, as well as the students’ learning paths.

ACTUAL OR ANTICIPATED OUTCOMES The first results obtained have made it possible to visualize the learning process, as well as the learning paths followed by each learner. They also provide analysis indicators for understanding and optimizing the learning process and the students’ paths in digital learning environments. These results allow the stakeholders (training managers, teachers, and students) to improve the way they teach and learn.

CONCLUSIONS/RECOMMENDATIONS/SUMMARY This approach made it possible to comprehensively understand the learning processes and the learning paths of each learner, to visualize their differences, as well as their advantages and disadvantages. The analysis of the learning processes promoted a correlation study between the behavior of the learner (i.e. the number of connections between the sections of a course followed) during the learning process and their mark obtained on the final exam. The correlation coefficient of the evaluated courses was of the order of 0.49 and 0.53 respectively. Moreover, in order to improve the predictive model, it's necessary to implement advanced analysis: diagnosis, predictive and prescriptive based on the descriptive elements (Process visualization). This allows teachers to have an integrated tool for analyzing learning traces through a monitoring, diagnostic, alert, and early intervention system in order to better promote the success of students.

KEYWORDS Learning Analytics, Learning Process, Process Mining, Learning Paths, Online Education