Cost-Bounded, Accuracy-Aware Hyperparameter Optimization Integrated with Rule-Based Book Recommendation for Developer Profiling on Stack Overflow
Informatica · 04.08.2026 · 50 (2)
DOI: 10.31449/inf.v50i2.11275 ↗Kurz gefasst
Fachartikel in der Zeitschrift Informatica, online erschienen am 04.08.2026. Verfasst von Fatma Altınsoy und 1 weiteren Autor:innen.
Abstract (Originalsprache)
This paper introduces HyperR, an integrated framework that combines hyperparameter optimization with a rule-based book recommender for developer profiling on Stack Overflow. The proposed optimizationmethod employs a cost-bounded, accuracy-aware strategy with early stopping and adaptive search to balance computational efficiency and search effectiveness. The framework evaluates both model-free (Grid Search, Random Search, Nelder–Mead, DEoptim) and model-based (Bayesian Optimization, AutoFT, mlrMBO,TuRBO) optimizers across multiple classifiers on a dataset of 20,000 Stack Overflow questions, using stratified 70/30 train-test splits with 20 repeated runs per configuration. Experimental results show that the proposed method achieves the highest macro-averaged precision (0.671) and significantly outperformsclassical baselines (adjusted p < 0.008), while remaining competitive with state-of-the-art modelbased approaches (adjusted p > 0.05). In addition, it reduces peak memory usage by up to 25% compared with exhaustive search methods. The recommendation module achieves strong performance (Precision@3= 0.82 and Recall@3 = 0.77), demonstrating the effectiveness of the proposed framework in delivering personalized learning resources.
Zitieren
APA
Altınsoy, F., & Öztürk, M. M. (2026). Cost-Bounded, Accuracy-Aware Hyperparameter Optimization Integrated with Rule-Based Book Recommendation for Developer Profiling on Stack Overflow. Informatica, 50(2). https://doi.org/10.31449/inf.v50i2.11275
Vancouver
Altınsoy F, Öztürk MM. Cost-Bounded, Accuracy-Aware Hyperparameter Optimization Integrated with Rule-Based Book Recommendation for Developer Profiling on Stack Overflow. Informatica. 2026;50(2). doi:10.31449/inf.v50i2.11275
BibTeX
@article{Altnsoy2026,
author = {Altınsoy, Fatma and Öztürk, Muhammed Maruf},
title = {Cost-Bounded, Accuracy-Aware Hyperparameter Optimization Integrated with Rule-Based Book Recommendation for Developer Profiling on Stack Overflow},
journal = {Informatica},
year = {2026},
volume = {50},
number = {2},
doi = {10.31449/inf.v50i2.11275}
}
EUVIMED