Penerapan Data mining untuk Prediksi Turnover Karyawan Rumah Sakit dalam Manajemen Sumber Daya Manusia Metode Hybrid Naïve Bayes dan C4.5
The high rate of employee turnover in hospitals, reaching 20-30% per year, has become a serious challenge for human resource management (HRM) because it affects productivity, operational costs, and the quality of health services. This research is grounded in data mining theory through the Knowledge Discovery in Database (KDD) framework and the Naïve Bayes and C4.5 (Decision tree) classification algorithms, which respectively offer probabilistic advantages and interpretable decision-tree capability in predicting employee turnover. This study employed a descriptive qualitative method with a Human Resource Analytics approach, utilizing secondary data from the Human Resource Analytics (HRIS) consisting of 153 employee records at a private hospital in Menteng, Central Jakarta. The research stages followed the KDD process, comprising data selection, preprocessing, transformation, data mining, and evaluation, with model testing carried out through a K-Fold Cross Validation scheme using RapidMiner. The results showed that the Naïve Bayes algorithm achieved an accuracy of 91.30% with an AUC of 0.971, while C4.5 produced a higher accuracy of 97.83% with an AUC of 0.993, so that C4.5 is recommended as the primary predictive model. In conclusion, the application of data mining with a hybrid Naïve Bayes and C4.5 approach proved effective and feasible as the basis for a data-driven decision support system in formulating hospital employee retention strategies
Authors
- Niko Suwaryo
- Safitri Hidayah
- Ag. Dwi Winarno
Institutions
- Universitas Medika Suherman (ID)
Publication Details
- Journal
- CEMERLANG Jurnal Manajemen dan Ekonomi Bisnis
- Published
- 2026-11-14
- DOI
- https://doi.org/10.55606/cemerlang.v6i4.10697
- Primary Topic
- AI and HR Technologies
- Type
- article
- Field-Weighted Citation Impact
- 0.00