Prediction of Experience Status Based on Human Resources Data: A Comparison of Lasso, Ridge, and Elastic Net Logistic Regression Models
VII. International Applied Statistics Congress (UYIK – 2026), İstanbul, Türkiye, 11 - 13 Mayıs 2026, ss.176, (Özet Bildiri)
- Yayın Türü: Bildiri / Özet Bildiri
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.176
- Ankara Üniversitesi Adresli: Evet
Özet
This study proposes a logistic regression–based modeling framework to predict employees’ experience status using human resources data. To improve model generalization and control for multicollinearity, Ridge, LASSO, and Elastic Net regularization techniques are comparatively examined. The models were trained on a dataset comprising employees’ demographic and professional characteristics, and their performance was evaluated using accuracy, precision, recall, F1-score, and ROC-AUC metrics. The findings indicate that all methods demonstrate strong classification performance and produce consistent results across the dataset. Among the approaches, LASSO produces simpler and more interpretable models, while Elastic Net provides greater flexibility in handling correlated predictors. Ridge regression, on the other hand, enhances model stability through coefficient shrinkage. Overall, incorporating regularization techniques into logistic regression proves to be an effective strategy for predictive analytics in human resources data. The study offers an analytical framework that can support decision-making processes by modeling the relationship between employee characteristics and experience status.