EUVIMEDEUVIMED entdecken
Crossref

A hybrid data-driven framework integrating machine learning and multi-criteria decision making for recruitment

Özlem Üstün, Merve Yilmaz, Hasan Şahin

Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi · 16.11.2026 · (Advanced Online Publication)

DOI: 10.65206/pajes.1988136 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi, online erschienen am 16.11.2026. Verfasst von Özlem Üstün und 2 weiteren Autor:innen. Die Arbeit nennt 75 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Context—Recruitment has become a multidimensional decision-making process in which organizations must evaluate both qualitative and quantitative candidate attributes simultaneously. While machine iearning (ML) methods are effective for predicting recruitment outcomes and Multi-Criteria Decision-Making (MCDM) techniques provide structured ranking mechanisms, these approaches are commonly applied independently. As a result, the advantages of predictive analytics and expert-based evaluation are not fully utilized within a unified decision-support processObjective—This study proposes a hybrid decision-support framework that integrates ML and MCDM techniques to improve personnel selection. The framework is designed to combine data-driven prediction with systematic multi-criteria evaluation, while also examining the relative importance of recruitment criteria and the consistency of alternative ranking methods.Method—The proposed methodology consists of two sequential stages. First, Decision Tree, Random Forest, and Gradient Boosting Tree algorithms are developed using a publicly available recruitment dataset, and their predictive performances are comparatively evaluated to identify the most suitable model. The predicted hiring probabilities are then used to shortlist the most promising candidates. In the second stage, these candidates are evaluated using an integrated MCDM framework. Criterion weights are determined through the Analytic Hierarchy Process (AHP), whereas candidate rankings are generated using TOPSIS, ELECTRE, and VIKOR. Finally, Spearman’s rank correlation coefficient is employed to examine the degree of agreement among the rankings obtained from the different MCDM methods.Results—Among the evaluated ML algorithms, Random Forest and Gradient Boosting Tree demonstrated closely comparable predictive performance. Gradient Boosting Tree was retained for candidate screening because it achieved slightly higher recall, F1-score, and test ROC–AUC. The integration of predicted hiring probabilities into the MCDM stage enabled a more focused evaluation of shortlisted candidates. TOPSIS, ELECTRE, and VIKOR consistently identified the same highest-ranked candidate, although varying degrees of agreement were observed across the remaining candidate rankings. Furthermore, the AHP results identified Interview (0.5031) and Personality (0.2316) as the most influential criteria, followed by Education (0.1481), Skill (0.0755), and Experience (0.0417).Conclusion—The proposed hybrid framework effectively combines predictive modeling with multi-criteria evaluation in a unified decision-support environment. By linking ML predictions with transparent ranking techniques, the framework improves the reliability and interpretability of recruitment decisions. Owing to its flexible structure, the proposed approach can be adapted to various industrial engineering applications and other decision-making problems involving multiple evaluation criteria.

Zitieren

APA

Üstün, Ö., Yilmaz, M., & Şahin, H. (2026). A hybrid data-driven framework integrating machine learning and multi-criteria decision making for recruitment. Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi. https://doi.org/10.65206/pajes.1988136

Vancouver

Üstün Ö, Yilmaz M, Şahin H. A hybrid data-driven framework integrating machine learning and multi-criteria decision making for recruitment. Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi. 2026. doi:10.65206/pajes.1988136

BibTeX
@article{stn2026,
  author = {Üstün, Özlem and Yilmaz, Merve and Şahin, Hasan},
  title = {A hybrid data-driven framework integrating machine learning and multi-criteria decision making for recruitment},
  journal = {Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi},
  year = {2026},
  number = {Advanced Online Publication},
  doi = {10.65206/pajes.1988136}
}
Kostenlos starten →