Banca de QUALIFICAÇÃO: ELVIS MEDEIROS DE MELO

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : ELVIS MEDEIROS DE MELO
DATE: 15/12/2021
TIME: 14:00
LOCAL: meet.google.com/koo-oybu-wwu
TITLE:
KEY WORDS:

Educational Data Science, Educational Process Mining, Graphs, Machine Learning, XAI.

PAGES: 88
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUMMARY:

When properly planned, it plays an important role in assessing the skills of everyone involved. In view of the reality of e-learning, the Federal University of Rio Grande do Norte institutionalized the platform Multiprova to support assessment processes at the institution. The task of understanding how students behave, identifying student profiles, and gaining insights through records of online assessment resolutions is a research field of Educational Process Mining, which can be performed from event log files with graph modeling. Each student's learning path becomes the object of study for modeling in this work, being described as the path taken by a student through a series of movements in an e-learning environment to carry out an assessment on the Multiprova platform. Furthermore, there is a need to transform interpretable Machine Learning models through the diffusion of eXplainable Artificial Intelligence (XAI) techniques. Based on this reality, the thesis hypothesis arises: Is it possible to use data from online test resolution records to gain insights into the learning process and student profiles in Multiprova using graph modeling and Machine Learning? Therefore, the theoretical foundation is presented on the themes that make up the object of study, such as Graphs and Analysis of Complex Networks, Machine Learning, XAI applied to Education, Educational Data Science and Educational Process Mining. When analyzing similar works in the literature, we observed that graph features are not used for the clustering process of students in EPM processes, much less use data from online assessment logs to perform such modeling, being case studies to solve specific problems. A chapter to introduce the Multiprova platform and the importance of assessment in the e-learning environment was built. A proof of concept modeling a Multiprova assessment was designed. Among the preliminary results, we realize the importance of using graph features in the clustering of students into three groups, applying data visualization techniques and XAI to interpret the results.

BANKING MEMBERS:
Presidente - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Externo ao Programa - 1555898 - DIEGO RODRIGO CABRAL SILVA
Externa ao Programa - 1362181 - ISMENIA BLAVATSKY DE MAGALHÃES