HELA: A novel hybrid ensemble learning algorithm for predicting academic performance of students

Education plays a major role in the development of the consciousness of the whole society. Education has been improved by analyzing educational data related to student academic performance. By using data mining techniques and algorithms on data from the educational environment, students' perfor...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Education and information technologies Ročník 27; číslo 4; s. 4521 - 4552
Hlavní autoři: Keser, Sinem Bozkurt, Aghalarova, Sevda
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York Springer US 01.05.2022
Springer
Springer Nature B.V
Témata:
Academic Achievement
Algorithms
Analysis
Computer Appl. in Social and Behavioral Sciences
Computer Science
Computers and Education
Data Analysis
Data Collection
Data mining
Education
Educational Environment
Educational Technology
Evaluation Methods
Grade Prediction
Information Systems Applications (incl.Internet)
Mathematics
Students
User Interfaces and Human Computer Interaction
Portugal
Ensemble learning algorithm
Prediction of academic performance
Educational data mining
Classification
ISSN:1360-2357, 1573-7608
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Education plays a major role in the development of the consciousness of the whole society. Education has been improved by analyzing educational data related to student academic performance. By using data mining techniques and algorithms on data from the educational environment, students' performances can be predicted. In this study, a novel Hybrid Ensemble Learning Algorithm (HELA) is proposed to predict the academic performance of students. The prediction results obtained from base classifiers namely Gradient Boosting, Extreme Gradient Boosting, Light Gradient Boosting Machine, and different combinations of these algorithms are given as input to the Super Learner algorithm. Hyper-parameters of base classifiers are optimized with a Random Search algorithm. Students' performances in Math and Portuguese classes are predicted by the proposed algorithm. In the experimental results, 96.6% and 91.2% accuracy values are obtained for the Mathematics course, and the Portuguese course, respectively. This paper is the first study, to our knowledge, to integrate the boosting and stacking-based ensemble learning algorithm for the prediction of students' academic performance that gives better predictive results with high efficiency.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1360-2357
1573-7608
DOI:10.1007/s10639-021-10780-0