Melola Discovery

← Melola Discovery

Akses terbuka Jurnal SINTA 1

Hybrid Search–Driven Hyperparameter Optimization to Enhance Dropout Prediction in Higher Education

Journal of Robotics and Control (JRC) · Vol. 7 No. 3 (2026) · hlm. 3894–3908

DOI
10.18196/jrc.v7i3.27423
ISSN
2715-5072, 2715-5056
Terbit
2026-05-01
Penerbit
Universitas Muhammadiyah Yogyakarta
Bahasa
eng
Lisensi
creativecommons.org/licenses/by-sa/4.0

Abstrak

Student dropout prediction remains a critical challenge in higher education, particularly for institutions with limited resources that require reliable early-warning mechanisms. This study addresses this problem by proposing a hybrid search–driven stacking ensemble designed to overcome the limitations of single learners in handling noisy, incomplete, and heterogeneous student data. The research contribution is threefold: (1) the development of a two-stage Random–Grid hybrid hyperparameter optimization strategy that improves search efficiency and predictive stability; (2) the formulation of a stacking architecture that integrates both bagging- and boosting-based base learners; and (3) empirical validation demonstrating the ensemble’s robustness under parameter perturbations and varied training folds. The Academic Success Dataset from the UCI Repository, consisting of 127,530 samples and 38 features, is utilized. After preprocessing, the data are split into 80% training and 20% testing, with the training portion evaluated using 10-fold cross-validation. Multiple ensemble models—Random Forest, XGBoost, LightGBM, CatBoost, and AdaBoost—are trained and optimized before being integrated into the proposed stacking framework. Experimental results show that the optimized stacking ensemble achieves 98.34% accuracy, 96.61% recall, 99.32% precision, 97.95% F1-score, and an AUC-ROC of 0.9992, outperforming all baseline models and recent state-of-the-art approaches. These findings confirm the model’s strong discriminative capability and its potential to enhance data-driven early-intervention systems. Overall, the proposed framework provides an efficient, scalable, and empirically validated solution for dropout risk detection in higher education.

Kata kunci

Jaringan kolaborasi para penulis

12 penulis · 26 tautan kepenulisan bersama · 10 lintas organisasi

Arahkan kursor atau ketuk sebuah nama untuk menyorot rekan penulisnya. Klik dua kali untuk membuka halaman penulis.

  • Universitas Muhammadiyah Riau
  • Universiti Malaysia Kelantan
  • Organisasi berbeda
  • Organisasi sama
  • Afiliasi tidak diketahui
  • Garis lebih tebal = lebih banyak artikel bersama · lingkaran lebih besar = lebih banyak artikel
Tampilkan sebagai tabel
Rekan penulisOrganisasiArtikel bersama
Vitriani VitrianiUniversitas Muhammadiyah Riau2
Rahmad GunawanUniversitas Muhammadiyah Riau2
Rahmad Al RianUniversitas Muhammadiyah Riau2
Hadhrami Ab GhaniUniversiti Malaysia Kelantan2
Pratama Benny HerlandyUniversitas Muhammadiyah Riau1
Diah Eka RatnaUniversitas Muhammadiyah Riau1
Hadhrami Ab GhaniUniversiti Malaysia Kelantan1
Nurul Izrin Md SalehUniversiti Malaysia Kelantan1

Afiliasi berasal dari OpenAlex; metadata OAI-PMH jurnal tidak memuatnya. Penulis tanpa ID OpenAlex dicocokkan berdasarkan nama, sehingga dua orang bisa tergabung atau satu orang bisa terpisah.

Sumber metadata: OAI-PMH (metadata jurnal) https://journal.umy.ac.id/index.php/jrc/oai, pertama di-harvest 9 Nov 2026, terakhir terlihat 16 Nov 2026.
Metadata asli (oai_dc)
<oai_dc:dc xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/  http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
	<dc:title xml:lang="en-US">Hybrid Search–Driven Hyperparameter Optimization to Enhance Dropout Prediction in Higher Education</dc:title>
	<dc:creator>Edi Ismanto</dc:creator>
	<dc:creator>Harun Mukhtar</dc:creator>
	<dc:creator>Wan Laura Hardilawati </dc:creator>
	<dc:creator>Hadhrami Ab Ghani</dc:creator>
	<dc:subject xml:lang="en-US">Hybrid Search Optimization</dc:subject>
	<dc:subject xml:lang="en-US">Meta-Learning Architecture</dc:subject>
	<dc:subject xml:lang="en-US">Ensemble Hyperparameter Tuning</dc:subject>
	<dc:subject xml:lang="en-US">Supervised Dropout Modeling</dc:subject>
	<dc:subject xml:lang="en-US">Performance–Cost Efficient Learning</dc:subject>
	<dc:description xml:lang="en-US">Student dropout prediction remains a critical challenge in higher education, particularly for institutions with limited resources that require reliable early-warning mechanisms. This study addresses this problem by proposing a hybrid search–driven stacking ensemble designed to overcome the limitations of single learners in handling noisy, incomplete, and heterogeneous student data. The research contribution is threefold: (1) the development of a two-stage Random–Grid hybrid hyperparameter optimization strategy that improves search efficiency and predictive stability; (2) the formulation of a stacking architecture that integrates both bagging- and boosting-based base learners; and (3) empirical validation demonstrating the ensemble’s robustness under parameter perturbations and varied training folds. The Academic Success Dataset from the UCI Repository, consisting of 127,530 samples and 38 features, is utilized. After preprocessing, the data are split into 80% training and 20% testing, with the training portion evaluated using 10-fold cross-validation. Multiple ensemble models—Random Forest, XGBoost, LightGBM, CatBoost, and AdaBoost—are trained and optimized before being integrated into the proposed stacking framework. Experimental results show that the optimized stacking ensemble achieves 98.34% accuracy, 96.61% recall, 99.32% precision, 97.95% F1-score, and an AUC-ROC of 0.9992, outperforming all baseline models and recent state-of-the-art approaches. These findings confirm the model’s strong discriminative capability and its potential to enhance data-driven early-intervention systems. Overall, the proposed framework provides an efficient, scalable, and empirically validated solution for dropout risk detection in higher education.</dc:description>
	<dc:publisher xml:lang="en-US">Universitas Muhammadiyah Yogyakarta</dc:publisher>
	<dc:date>2026-05-01</dc:date>
	<dc:type>info:eu-repo/semantics/article</dc:type>
	<dc:type>info:eu-repo/semantics/publishedVersion</dc:type>
	<dc:format>application/pdf</dc:format>
	<dc:identifier>https://journal.umy.ac.id/index.php/jrc/article/view/27423</dc:identifier>
	<dc:identifier>10.18196/jrc.v7i3.27423</dc:identifier>
	<dc:source xml:lang="en-US">Journal of Robotics and Control (JRC); Vol. 7 No. 3 (2026); 3894-3908</dc:source>
	<dc:source>2715-5072</dc:source>
	<dc:source>2715-5056</dc:source>
	<dc:language>eng</dc:language>
	<dc:relation>https://journal.umy.ac.id/index.php/jrc/article/view/27423/12680</dc:relation>
	<dc:rights xml:lang="en-US">Copyright (c) 2026 Edi Ismanto, Harun Mukhtar, Wan Laura Hardilawati , Hadhrami Ab Ghani</dc:rights>
	<dc:rights xml:lang="en-US">https://creativecommons.org/licenses/by-sa/4.0</dc:rights>
</oai_dc:dc>