Naslov
Meta-features of learning management system data

Autori
Oreški, Dijana

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of the Twelfth International Conference on e-Learning / Domazet, Bojana ; Raspopović Milić, Miroslava - Beograd : Metropolitan University, 2021, 40-44

Skup
12th International Conference on e-Learning

Mjesto i datum
Beograd, Srbija, 23.09.2021. - 24.09.2021

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
meta-features ; LMS data ; machine learning ; meta-learning.

Sažetak
Educational institutions tend to use learning management systems (LMS) to enhance teaching and learning process. An LMS collects data about students’ activities in log files. This data is used in prediction of student success. Machine learning algorithms are increasingly used to support development of accurate and reliable descriptive and predictive models from educational data. However, selection of machine learning algorithm is complex task and depends on characteristics of dataset used in the analysis. Characteristics of dataset (meta-features) can explain why one machine learning algorithm performs better on certain dataset than other algoriths. Meta-learning field deals with this issue. Identification of meta-feature in educational domain is still insufficiently researched area. In this paper, we are trying to fill that gap by providing meta-features identification in educational domain. Our focus are general meta-features and learning management system data. Research results indicated patterns in LMS datasets meta-features.

Izvorni jezik
Engleski

Znanstvena područja
Informacijske i komunikacijske znanosti

POVEZANOST RADA