Önerilen Yapay Sinir Ağı Algoritması ile Ortaokul Öğrencilerin Akademik Performansının Tahmini

Öz

Eğitsel veri madenciliği, eğitim sürecine ilişkin elde edilen büyük veri üzerinde farklı kaynakları kullanarak şimdiki zamana ve gelecek zamana ilişkin tahmin yapmamızı sağlayacak kural ve ilişkileri araştırır. Eğitsel veri madenciliği ile veri madenciliği alanındaki teknik ve algoritmaların kullanılmasıyla öğrenci veya eğitmenlerin akademik performansları tahmin edilebilir. Bu çalışmada, orta okul öğrencilerinin akademik performansını tahmin etmek amacıyla yeni bir yapay sinir ağı algoritması önerilmektedir. Önerilen algoritma, öncelikli olarak dengesiz sınıf dağılımı problemini çözmek için aşırı örnekleme tekniklerinden SMOTE algoritmasını önişleme aşamasında uygulamaktadır. Daha sonra, öznitelik seçim ve veri normalizasyonu işlemleri yapılarak çalışılan öğrenci veri seti, önerilen algoritmanın kullanımına hazır hale getirilmektedir. Çalışmada kullanılan rastgele arama algoritması ile yapay sinir ağı modelinin hiper-parametreleri optimize edilmektedir. Öğrencilerin Matematik ve Portekizce derslerindeki başarıları 2-seviyeli ve 5-seviyeli sınıflandırma için önerilen algoritma ile tahmin edilmektedir. Deney sonuçlarında, Matematik dersi için %97.0 ve %92.3 doğruluk değerleri sırasıyla 2-seviyeli ve 5-seviyeli sınıflandırma için elde edilmektedir. Protekizce dersi için ise bu değerler sırasıyle %97.6 ve %87.9 olarak hesaplanmaktadır.

Anahtar Kelimeler

• Eğitsel Veri Madenciliği
• Yapay Sinir Ağları
• Rastgele Arama

Prediction of Secondary School Students' Academic Performance with Proposed Artificial Neural Networks Algorithm

Öz

Educational data mining explores the rules and relationships that will enable us to make predictions about the present and the future by using different sources on the big data obtained about the educational process. With educational data mining, students' or instructors’ academic performance is predicted by using various techniques and algorithms in data mining. In this study, a new artificial neural network algorithm is proposed to predict the academic performance of secondary school students. The proposed algorithm primarily applies the SMOTE algorithm, one of the oversampling techniques, in the preprocessing stage to solve the problem of unbalanced class distributions. Then, the student dataset is made ready for the use of the proposed algorithm by performing feature selection and data normalization processes. The hyper-parameters of the artificial neural network model are optimized by using the random search algorithm used in the study. Students' achievement in Mathematics and Portuguese lessons are predicted for 2-level and 5-level classification. In the experimental results, the 97.0% and 92.3% accuracy values for the Mathematics course are obtained for 2-level and 5-level classification, respectively by using the proposed algorithm. For the Portuguese course, these values are calculated as 97.6% and 87.9%, respectively.

Anahtar Kelimeler

• Educational Data Mining
• Artificial Neural Networks
• Random Search

Kaynakça

1. Abdullah, A., BEŞERİ SERMAYENİN KALKINMA ÜZERİNE ETKİSİ. Uluslararası Ekonomi Siyaset İnsan ve Toplum Bilimleri Dergisi. 1(1): p. 28-34.
2. Kayadibi, F., Eğitim kalitesine etki eden faktörler ve kaliteli eğitimin üretime katkısı. İstanbul Üniversitesi İlahiyat Fakültesi Dergisi, 2001(3).
3. Romero, C. and S. Ventura, Educational data mining and learning analytics: An updated survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 2020. 10(3): p. e1355.
4. Salloum, S.A., et al. Mining in Educational Data: Review and Future Directions. in Joint European-US Workshop on Applications of Invariance in Computer Vision. 2020. Springer.
5. Peña-Ayala, A., Educational data mining: A survey and a data mining-based analysis of recent works. Expert systems with applications, 2014. 41(4): p. 1432-1462.
6. Shahiri, A.M. and W. Husain, A review on predicting student's performance using data mining techniques. Procedia Computer Science, 2015. 72: p. 414-422.
7. Satyanarayana, A. and M. Nuckowski, Data mining using ensemble classifiers for improved prediction of student academic performance. 2016.
8. Chaudhury, P., et al. Enhancing the capabilities of student result prediction system. in Proceedings of the Second International Conference on Information and Communication Technology for Competitive Strategies. 2016.

9. Salal, Y., S. Abdullaev, and M. Kumar, Educational Data Mining: Student Performance Prediction in Academic. IJ of Engineering and Advanced Tech, 2019. 8(4C): p. 54-59.
10. Hamoud, A., Selection of best decision tree algorithm for prediction and classification of students’ action. American International Journal of Research in Science, Technology, Engineering & Mathematics, 2016. 16(1): p. 26-32.
11. Pojon, M., Using machine learning to predict student performance. 2017.
12. Başer, S.H., O. Hökelekli, and A. Kemal, Ortaöğretimde Öğrenim Gören Öğrenci Performanslarının Veri Madenciliği Yöntemleri İle Tahmin Edilmesi. Bilgisayar Bilimleri ve Teknolojileri Dergisi. 1(1): p. 22-27.
13. Ünal, F., Data Mining for Student Performance Prediction in Education, in Data Mining-Methods, Applications and Systems. 2020, IntechOpen.
14. Athani, S.S., et al. Student academic performance and social behavior predictor using data mining techniques. in 2017 International Conference on Computing, Communication and Automation (ICCCA). 2017. IEEE.
15. Ma, X. and Z. Zhou. Student pass rates prediction using optimized support vector machine and decision tree. in 2018 IEEE 8th Annual Computing and Communication Workshop and Conference (CCWC). 2018. IEEE.
16. Troussas, C., M. Virvou, and S. Mesaretzidis. Comparative analysis of algorithms for student characteristics classification using a methodological framework. in 2015 6th International Conference on Information, Intelligence, Systems and Applications (IISA). 2015. IEEE.
17. Singh, M., et al. Towards enthusiasm prediction of Portuguese school's students towards higher education in realtime. in 2020 International Conference on Computation, Automation and Knowledge Management (ICCAKM). 2020. IEEE.
18. Walia, N., et al., Student’s Academic Performance Prediction in Academic using Data Mining Techniques. Available at SSRN 3565874, 2020.
19. Srivastava, A.K., et al., PREDICTION OF STUDENTS PERFORMANCE USING KNN AND DECISION TREE-A MACHINE LEARNING APPROACH. 2020.
20. Zaffar, M., et al., Role of FCBF Feature Selection in Educational Data Mining. Mehran University Research Journal of Engineering and Technology, 2020. 39(4): p. 772-778.
21. Haykin, S.S., Neural networks and learning machines/Simon Haykin. 2009, New York: Prentice Hall.
22. Ataseven, B., Yapay sinir ağları ile öngörü modellemesi. 2013.
23. Öztemel, E., Yapay sinir ağlari. PapatyaYayincilik, Istanbul, 2003.
24. Tran, N., et al., Hyper-parameter optimization in classification: To-do or not-to-do. Pattern Recognition, 2020. 103: p. 107245.
25. Young, S.R., et al. Optimizing deep learning hyper-parameters through an evolutionary algorithm. in Proceedings of the Workshop on Machine Learning in High-Performance Computing Environments. 2015.
26. Syarif, I., A. Prugel-Bennett, and G. Wills, SVM parameter optimization using grid search and genetic algorithm to improve classification performance. Telkomnika, 2016. 14(4): p. 1502.
27. Cortez, P. and A.M.G. Silva, Using data mining to predict secondary school student performance. 2008.
28. Bergstra, J. and Y. Bengio, Random search for hyper-parameter optimization. The Journal of Machine Learning Research, 2012. 13(1): p. 281-305.
29. Erten, G.E., C.V. Deutsch, and M. Yavuz, Managing Estimation Artifacts in Machine Learning Spatial Estimation.
30. BUDAK, H., Özellik Seçim Yöntemleri ve Yeni Bir Yaklaşım. Journal of Natural & Applied Sciences, 2018.
31. Maxwell, A., et al., Deep learning architectures for multi-label classification of intelligent health risk prediction. BMC bioinformatics, 2017. 18(14): p. 523.
32. Ohsaki, M., et al., Confusion-matrix-based kernel logistic regression for imbalanced data classification. IEEE Transactions on Knowledge and Data Engineering, 2017. 29(9): p. 1806-1819.
33. Cabena, P., et al., Discovering data mining: from concept to implementation. 1998: Prentice-Hall, Inc.
34. Gazel, S. and C.T. BATİ, Derin Sinir Ağları ile En İyi Modelin Belirlenmesi: Mantar Verileri Üzerine Keras Uygulaması. Yüzüncü Yıl Üniversitesi Tarım Bilimleri Dergisi. 29(3): p. 406-417.
35. Yam, J.Y. and T.W. Chow, A weight initialization method for improving training speed in feedforward neural network. Neurocomputing, 2000. 30(1-4): p. 219-232.