Browsing by Author "Sevgin, Hikmet"

Now showing 1 - 2 of 2

A Comparative Study of Ensemble Methods in the Field of Education: Bagging and Boosting Algorithms
(Izzet Kara, 2023) Sevgin, Hikmet
This study aims to conduct a comparative study of Bagging and Boosting algorithms among ensemble methods and to compare the classification performance of TreeNet and Random Forest methods using these algorithms on the data extracted from ABIDE application in education. The main factor in choosing them for analyses is that they are Ensemble methods combining decision trees via Bagging and Boosting algorithms and creating a single outcome by combining the outputs obtained from each of them. The data set consists of mathematics scores of ABIDE (Academic Skills Monitoring and Evaluation) 2016 implementation and various demographic variables regarding students. The study group involves 5000 students randomly recruited. On the deletion of loss data and assignment procedures, this number decreased to 4568. The analyses showed that the TreeNet method performed more successfully in terms of classification accuracy, sensitivity, F1-score and AUC value based on sample size, and the Random Forest method on specificity and accuracy. It can be alleged that the TreeNet method is more successful in all numerical estimation error rates for each sample size by producing lower values compared to the Random Forest method. When comparing both analysis methods based on ABIDE data, considering all the conditions, including sample size, cross validity and performance criteria following the analyses, TreeNet can be said to exhibit higher classification performance than Random Forest. Unlike a single classifier or predictive method, the classification or prediction of multiple methods by using Boosting and Bagging algorithms is considered important for the results obtained in education.
Comparison of Classification Performances of Mars and Brt Data Methods: Ab?de-2016 Case
(Turkish Education Assoc, 2022) Sevgin, Hikmet; Onen, Emine
This research examined the relationships between student, teacher, school and instructional qualifications and 8th grade students' science achievement, based on the conceptual framework created by Nilsen and Gustafsson (2016), using data mining methods MARS and BRT. Research data (n=10407 students, n=941 teachers and n=865 school administrators) were obtained from the AB??DE study conducted at the national level by the Ministry of National Education in 2016. MARS and BRT analyzes were performed in the SPM 8.2 program. The science achievement classification performances of these methods were compared by considering the correct classification rate, sensitivity and specificity rates, F1 statistical value and the area under the ROC curve. It was found that the BRT method was more successful than the MARS method in terms of all these criteria, and the most important predictors of science achievement were similar compared to these two methods. The results revealed that the most important predictors of science success are the student's perception of science self-efficacy, the father's occupation, the family's monthly income, the instructional activities of the teacher, the teacher's experience and preparation for the lesson, and the school administrators' perception of school climate. It is thought that the reason why BRT outperforms the MARS method in terms of the criteria considered in this study is that BRT learns from errors with the additive combination of various regression trees and provides a stronger classification performance by minimizing the errors that may occur in classification. This study revealed the benefits of using these two data mining methods in the field of Educational Sciences and discussed the contribution of the related methods in this field.