Predictive Analytics in Human Resources Management: Evaluating AIHR’s Role in Talent Retention
Abstract
1. Introduction
- Firstly, based on a PRISMA adapted methodology we selected the most impressive works that emphasize machine learning-based techniques applied to human resource management.
- Secondly, we identify and compare AI algorithms used in HR applications to underline the most robust ones for this topic and to discover in this way the gaps where new hybrid solutions can be further developed for a robust AIHR approach.
2. Materials and Methods
- Articles that use the most advanced analysis techniques.
- Articles written in English.
- Articles published in the last four years (2019–2024).
- Duplicate references from different electronic archives searched.
- Articles that did not address any of the research questions.
- Articles that did not align with the study objectives.
- Articles whose abstracts did not match the keywords from the search criteria.
3. Context of Management Theory Framework
4. Introducing Artificial Intelligence in Human Resources Analytics
- A higher salary or a more attractive role offered elsewhere,
- An unsuitable or uncomfortable work environment,
- Misalignment with the company’s goals,
- A lack of work–life balance, leading to stress,
- Relocation or pursuit of higher education opportunities,
- A lack of appreciation or feeling underutilized,
- Excessive overtime,
- Observing the departure of talented employees or team leaders from the organization.
5. AI-Based Solution Development for Human Resource Management
5.1. Research Findings in HR Explorations
- Gaussian Naive Bayes;
- Multinomial Naive Bayes;
- Bernoulli Naive Bayes;
- Decision trees;
- Random forest;
- Logistic regression.
5.2. Predictive Approaches
- ADASYN: This creates synthetic instances for the minority class, thereby increasing the proportion of data for employees leaving the company and improving algorithm performance in predictions for this class.
- Undersampling: This reduces the number of observations in the majority class, retaining only a portion to balance the data and achieve high performance in situations with imbalanced datasets.
6. Comparative Analysis
6.1. Critical Model Performance Analysis
6.2. Answers to the Research Questions
- Recruitment and selection: AI algorithms assist in sorting candidates and reducing biases. Machine learning models improve the match between candidates and roles.
- Performance management: AI provides behavior analysis for more accurate evaluations.
- Turnover prediction: Predictive models help HR identify employees at risk of leaving and intervene proactively.
- XGBoost and random forest: Proven effective in turnover predictions due to their ability to handle large and complex datasets. XGBoost demonstrated excellent performance with an AUC of 0.86 in general studies and 0.95 in a specific context.
- Logistic regression and SVM: Delivered weaker results in comparison but have the advantage of easy interpretability.
- Neural networks and ensemble models: Achieved the best results for complex predictions, with increased accuracy in studies on large datasets.
- One-hot encoding and data standardization ensure model compatibility.
- Feature selection based on Pearson correlation reduces redundancy and increases model efficiency.
- Balancing algorithms such as ADASYN and undersampling assist in handling imbalanced datasets.
- Ensemble boosting and random forest achieve excellent results on large and complex datasets.
- Linear regression and logistic regression are effective for simpler datasets but are limited in handling more complex predictions.
- Decision trees and SVM show variable results but require advanced preprocessing for high performance.
7. Future Perspectives
- Antoniuk, Ivens, and Kolyada (2025) [65] examine how artificial intelligence reshapes the HR function through a model of adaptive management, where career development decisions are guided by predictive models and contextual factors. Their study introduces a framework that enables organizations to personalize interventions for each employee by considering performance, feedback, and preferences, thus promoting fluid and individualized career management. The authors highlight the benefits of human–machine strategic collaboration in decision-making processes.
- In another study, Hoa Do, Lin Xiao Chu, and Helen Shipton (2025) propose an AI-driven HRM model aimed at supporting employee resilience. It integrates real-time feedback mechanisms, behavioral monitoring, and adaptive recommendations for training and internal mobility. The model emphasizes AI’s capability to detect early signs of burnout, disengagement, or professional stagnation and respond proactively through personalized suggestions. Contextual data and psychosocial indicators are used to continuously adjust retention and development strategies [66].
- Mullens and Shen (2025) introduce the 2ACT (AI-Accentuated Career Transitions) framework, which uses AI to build skill bridges between employees’ current positions and future career paths. By dynamically analyzing skill gaps and labor market trends, 2ACT generates suggestions for internal or external transitions, supported by personalized learning paths. The model combines natural language processing (NLP) for user interaction with recommendation algorithms to propose concrete steps for career progression [64].
8. Conclusions
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
Abbreviations
AI | Artificial intelligence |
HRM | Human resource management |
HR | Human resource |
ML | Machine learning |
SVM | Support vector machine |
KNN | K-nearest neighbors |
AUC | Area under the curve |
NLP | Natural language processing |
RPA | Robotic process automation |
DT | Decision tree |
XAI | Explainable artificial intelligence |
PRISMA | Preferred Reporting Items for Systematic Reviews and Meta-Analyses |
AIHR | Analytics in human resources management |
EWS | Early warning system |
RAG | Red, amber, green |
EACA | Expected attrition cost after the retention period |
EACB | Expected attrition cost before the retention period |
NB | Naïve Bayes |
LR | Logistic regression |
RF | Random forest |
GBM | Gradient boosting machine |
GBDT | Gradient boosted decision trees |
ODBC | Open database connectivity |
API | Application programming interface |
ADASYN | Adaptive synthetic sampling |
LIME | Local interpretable model-agnostic explanations |
SHAP | Shapley addictive explanations |
References
- Ghedabna, L.; Ghedabna, R.; Imtiaz, Q.; Faheem, M.A.; Alkhayyat, A.; Hosen, M.S. Artificial Intelligence in Human Resource Management: Revolutionizing Recruitment, Performance, and Employee Development. Nanotechnol. Percept. 2024, 20, 52–68. [Google Scholar]
- Allen, D.G.; Bryant, P.C.; Vardaman, J.M. Retaining talent: Replacing misconceptions with evidence-based strategies. Acad. Manag. Perspect. 2010, 24, 48–64. [Google Scholar]
- Meyer, J.P.; Allen, N.J. A three-component conceptualization of organizational commitment. Hum. Resour. Manag. Rev. 1991, 1, 61–89. [Google Scholar] [CrossRef]
- Hom, P.W.; Lee, T.W.; Shaw, J.D.; Hausknecht, J.P. One hundred years of employee turnover theory and research. J. Appl. Psychol. 2017, 102, 530–545. [Google Scholar] [CrossRef]
- Lee, T.W.; Mitchell, T.R.; Holtom, B.C.; McDaniel, L.S.; Hill, J.W. The unfolding model of voluntary turnover: A replication and extension. Acad. Manag. J. 1999, 42, 450–462. [Google Scholar] [CrossRef]
- DeNisi, S.; Pritchard, A. Performance appraisal, performance management and improving individual performance: A motivational framework. Manag. Organ. Rev. 2006, 2, 253–277. [Google Scholar] [CrossRef]
- de Vreede, T.; Leidner, D. How AI is Making Self-Appraisals Better and Less Painful—For Both Workers and Managers, Business Insider. Available online: https://www.businessinsider.com/ai-improve-self-appraisals-performance-reviews-women-careers-promotions-2025-4 (accessed on 13 April 2025).
- AI vs. ML: What’s the Difference? Geek Culture. Available online: https://medium.com/geekculture/ai-vs-ml-whats-the-difference-2e2e73e98d9c (accessed on 20 November 2024).
- Adams, J.S. Inequity in social exchange. In Advances in Experimental Social Psychology; Berkowitz, L., Ed.; Academic Press: New York, NY, USA, 1965; Volume 2, pp. 267–299. [Google Scholar]
- Page, M.; McKenzie, J. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. In Proceedings of the RESEARCH METHODS AND REPORTING, UK, 21 March 2021; pp. 1–7. [Google Scholar]
- Wziątek, A.; Michalik, I.; Vveinhardt, J. Organizational commitment in the assessment of employees of different generations: A research study. Eur. Res. Stud. J. 2023, XXVI, 534–555. [Google Scholar] [CrossRef]
- Todorović, D.; Mitić, P.M.; Stojiljković, N.; Olanescu, M.; Suciu, A.; Popa, D. Organizational commitment in the private and public sectors. Front. Psychol. 2024, 15, 1442990. [Google Scholar] [CrossRef] [PubMed]
- Nosratabadi, S.; Mosavi, N.; Tan, K.J. Artificial intelligence models and employee lifecycle management: A systematic literature review. Technol. Forecast. Soc. Change 2022, 185, 122033. [Google Scholar] [CrossRef]
- Prasad, K.D.V.; De, T. Generative AI as a catalyst for HRM practices: Mediating effects of trust. HRM Rev. 2024, 17, 1362. [Google Scholar] [CrossRef]
- Vroom, V.H. Work and Motivation; Wiley: New York, NY, USA, 1964. [Google Scholar]
- Chakraborty, R.; Mridha, K.; Shaw, R.N.; Ghosh, A. Study and Prediction Analysis of the Employee Turnover using Machine Learning Approaches. In Proceedings of the 2021 IEEE 4th International Conference on Computing, Power and Communication Technologies (GUCON), Kuala Lumpur, Malaysia, 24–26 September 2021. [Google Scholar]
- Murire, O.T. Artificial Intelligence and Its Role in Shaping Organizational Work Practices and Culture. Adm. Sci. 2024, 14, 316. [Google Scholar] [CrossRef]
- Gayen, S. New Age HR, MANAV 2023. Available online: https://www.iiswbm.edu/wp-content/uploads/2022/08/departmnetal-activities-mba-hrm.pdf (accessed on 5 January 2025).
- Geetha, R.; Bhanu, S.R.D. Recruitment Through Artificial Intelligence: A Conceptual Study. Int. J. Mech. Eng. Technol. 2018, 9, 63–70. [Google Scholar]
- Bahukhandi, S.P.M. Applicants’ perception towards the application of AI in recruitment process. In Proceedings of the 2022 Interdisciplinary Research in Technology and Management (IRTM), Kolkata, India, 24–26 February 2022. [Google Scholar]
- Achchab, S.; Temsamani, Y.K. Artificial Intelligence Use in Human Resources Management: Strategy and Operation’s Impact. In Proceedings of the 2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning, Chengdu, China, 16–18 July 2021. [Google Scholar]
- Lin, H.M.Z.; Xu, Y.; Jiang, K.; Ma, W. A Muti-Field Feature Interaction Convolutional Neural Network for Resume Recommendation. In Proceedings of the 2020 International Symposium on Autonomous Systems (ISAS), Guangzhou, China, 6–8 September 2020. [Google Scholar]
- Chaudhary, A.; Jobanputra, M.; Shah, S.; Gandhi, R.; Chaudhary, S.; Goswamim, R. Automated Human Capital Management System. In Proceedings of the 2018 Annual IEEE International Systems Conference (SysCon) 2018 IEEE, Vancouver, BC, Canada, 23–26 April 2018. [Google Scholar]
- Tewari, I.; Pant, M. Artificial Intelligence Reshaping Human Resource Management: A Review. In Proceedings of the IEEE International Conference on Advent Trends in Multidisciplinary Research and Innovation, Buldhana, India, 30 December 2020; pp. 1–4. [Google Scholar]
- Artelt, A.; Gregoriades, A. “How to make them stay?”—Diverse Counterfactual Explanations of Employee Attrition. In Proceedings of the International Conference on Enterprise Information Systems, ICEIS 2023, Prague, Czech Republic, 24–26 April 2023; Volume 1, p. 6. [Google Scholar]
- Arthur, J.F.; Mazor, H. The Employee Experience: Culture, Engagement, and Beyond, 2017 Global Human Capital Trends. Available online: https://www.deloitte.com/us/en/insights/topics/talent/human-capital-trends/2017/improving-the-employee-experience-culture-engagement.html (accessed on 30 January 2025).
- What Will HR Focus on in 2023? Available online: https://www.gartner.com/en/articles/what-will-hr-focus-on-in-2023 (accessed on 20 November 2024).
- Zawacki-Richter, O.; Marín, V.I.; Bond, M.; Gouverneur, F. Systematic review of research on artificial intelligence applications in higher education—Where are the educators? Int. J. Educ. Technol. High. Educ. 2019, 16, 1–27. [Google Scholar] [CrossRef]
- Minbaeva, D. HRM in the era of AI: Challenges and opportunities. Hum. Resour. Manag. Rev. 2021, 31, 100765. [Google Scholar]
- Tursunbayeva, A.; Di Lauro, S.; Pagliari, C. People analytics—A new era for HRM? J. Organ. Eff. 2018, 5, 201–217. [Google Scholar]
- Sharma, G. A literature Review on Application of Artificial Intelligence in Human Resource Management and Its Practices in Current Organizational Scenario. In Proceedings of the Fifth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC), Palladam, India, 11–13 November 2021; p. 7. [Google Scholar]
- Tursunbayeva, A. People analytics research and practice: A review of international conferences. Strateg. HR Rev. 2020, 19, 267–272. [Google Scholar] [CrossRef]
- Gupta, R. Impact of Artificial Intelligence (AI) on Human Resource Management (HRM). Int. J. Multidiscip. Res. 2024, 6, 10. [Google Scholar]
- Mourad, Z.; Noura, A.; Mohamed, C. Towards a new method for classifying employee performance using machine learning algorithms. In Proceedings of the 2022 International Conference on Intelligent Systems and Computer Vision (ISCV), Fez, Morocco, 18–20 May 2022. [Google Scholar]
- Kumar, M.R.; Bhargav, Y.K.; Sharma, A.; Ramesh, G. Human Resource Management Using Machine Learning-Based Solutions. In Proceedings of the Third International Conference on Electronics and Sustainable Communication Systems (ICESC 2022), Coimbatore, India, 17–19 August 2022; IEEE: Piscataway, NJ, USA, 2022. [Google Scholar]
- Key HR Statistics and Trends in 2023–Forbes Advisor. Available online: https://www.forbes.com/advisor/business/hr-statistics-trends/ (accessed on 17 May 2023).
- Hausknecht, J.P.; Rodda, J.; Howard, M.J. Targeted employee retention: Performance-based and job-related differences in reported reasons for staying. Hum. Resour. Manag. 2009, 48, 269–288. [Google Scholar] [CrossRef]
- Abonamah, A.; La Torre, D.; Poulin, M.; Repetto, M. Explainable Artificial Intelligence in Human Resources: A Computational Study. In Proceedings of the 2022 International Conference on Data Analytics for Business and Industry (ICDABI), Sakhir, Bahrain, 25–26 October 2022. [Google Scholar]
- Sadana, P.; Munnuru, D. Machine Learning Model to Predict Work Force Attrition. In Proceedings of the 2021 6th International Conference for Convergence in Technology (I2CT), Pune, India, 2–4 April 2021. [Google Scholar]
- AI in HR: How AI Is Transforming the Future of HR|Gartner. 2023. Available online: https://www.gartner.com/en/human-resources/topics/artificial-intelligence-in-hr (accessed on 30 January 2025).
- Krishna, S.; Sidharth, S. Analyzing Employee Attrition Using Machine Learning: The New AI Approach. In Proceedings of the 2022 IEEE 7th International Conference for Convergence in Technology (I2CT), Pune, India, 7–9 April 2022. [Google Scholar]
- Darapaneni, N.; Turaga, R.N.; Shah, V.C.; Paduri, A.R.; Kumar, R.D.; Suram, M.; Venkatraman, V. A Detailed Analysis of AI Models for Predicting Employee Attrition Risk. In Proceedings of the 2022 IEEE 10th Region 10 Humanitarian Technology Conference (R10-HTC), Hyderabad, India, 16–18 September 2022. [Google Scholar]
- Mhatre, A.; Mahalingam, A.; Narayanan, M.; Nair, A.; Jaju, S. Predicting Employee Attrition along with Identifying High Risk Employees using Big Data and Machine Learning. In Proceedings of the 2020 2nd International Conference on Advances in Computing, Communication Control and Networking, Greater Noida, India, 18–19 December 2020. [Google Scholar]
- The Great Employment Turnover Statistics in 2023|Shortlister. Available online: https://www.myshortlister.com/insights/employment-turnover-statistics (accessed on 30 January 2025).
- Yuan, J. Research on Employee Performance Prediction Based on Machine Learning. In Proceedings of the 2022 IEEE 5th International Conference on Electronics Technology (ICET), Chengdu, China, 13–16 May 2022. [Google Scholar]
- Tarambiwa, V.M.E.; Zvarevashe, K. A Supervised Machine Learning Model to Optimize Human Resources Analytics for Employee Churn Prediction. In Proceedings of the Zimbabwe Conference on Information Communication & Technology (ZCICT), Harare, Zimbabwe, 9–10 November 2022. [Google Scholar]
- Habous, A.; Nfaoui, E.H.; Oubenaalla, Y. Predicting Employee Attrition using Supervised Learning Classification Model. In Proceedings of the 2021 IEEE 5th International Conference on Intelligence Computing in Data Sciences (ICDS), Fez, Morocco, 20–22 October 2021. [Google Scholar]
- HR_comma_sep.csv. Available online: https://www.kaggle.com/datasets/liujiaqi/hr-comma-sepcsv (accessed on 21 February 2025).
- Shi, Y. A Machine Learning Study on the Model Performance of Human Resources Predictive Algorithms. In Proceedings of the International Conference on Applied Machine Learning, Changsha, China, 23–25 July 2022. [Google Scholar]
- Kaur, B.; Dogra, A. A Machine Learning Model for Predicting Employees Retention: An Initiative towards HR through Machine. In Proceedings of the Seventh International Conference on Parallel, Distributed and Grid Computing (PDGC), Solan, India, 25–27 November 2022; p. 5. [Google Scholar]
- Costa, R.; Carvalgo, P. Machine Learning Predictive Models for preventing Employee Turnover cost. In Proceedings of the 2022 International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCME), Maldives, Maldives, 16–18 November 2022. [Google Scholar]
- Pandey, S.; Mehta, J. HRPA: Human Resource Prediction Analytics. In Proceedings of the 2022 Third International Conference on Intelligent Computing, Instrumentation and Control Technologies (ICICICT), Kannur, Keral, 11–12 August 2022. [Google Scholar]
- Sarah, S.; Alduayj, K.R. Predicting Employee Attrition using Machine Learning. In Proceedings of the International Conference on Innovations in Information Technology, 2018 13th IIT, Al Ain, United Arab Emirates, 18–19 November 2018. [Google Scholar]
- Zhang, H.; Xu, L.; Cheng, X.; Chao, K.; Zhao, X. Analysis and prediction of employee turnover characteristics based on machine learning. In Proceedings of the 18th International Symposium on Communications and Information Technologies (ISCIT 2018), Bangkok, Thailand, 26–28 September 2018. [Google Scholar]
- Mwaro, P.N.; Ogada, K.; Cheruiyot, W. Neural Network Model for Talent Recruitment and Management for Employee Development and Retention. In Proceedings of the 2021 IEEE AFRICON, Arusha, Tanzania, 13–15 September 2021. [Google Scholar]
- Joloudari, J.H.; Marefat, A.; Nematollahi, M.A.; Oyelere, S.S.; Hussain, S. Effective Class-Imbalance Learning Based on SMOTE and Convolutional Neural Networks. Appl. Sci. 2023, 13, 4006. [Google Scholar] [CrossRef]
- Raza, A.; Munir, K.; Almutairi, M.; Younas, F.; Fareed, M.M.S. Predicting employee attrition using machine learning approaches. Appl. Sci. 2022, 12, 6424. [Google Scholar] [CrossRef]
- Dsouza, P.K. Absolute answerability in the Era of Artificial Intelligence and Machine Learning: A talent management theperspective. In Proceedings of the 2019 International Conference on Digitization (ICD) IEEE 2019, Sharjah, United Arab Emirates, 18–19 November 2019. [Google Scholar]
- Yassminth, R. Time Complexity in ML: Ensuring Model Performance with Growing Data Sizes|by Rayan Yassminh|Medium. 2024. Available online: https://medium.com/@ryassminh/time-complexity-in-ml-ensuring-model-performance-with-growing-data-sizes-42007e5b7305 (accessed on 15 June 2025).
- Duan, J.; Wu, S. Beyond traditional pathways: Leveraging generative AI for dynamic career planning in vocational education. Int. J. New Dev. Educ. 2024, 6, 24–31. [Google Scholar] [CrossRef]
- Ma, X.; Liu, W.; Zhao, C.; Tukhvatulina, L.R. Can Large Language Model Predict Employee Attrition? arXiv 2024, arXiv:2411.01353. [Google Scholar] [CrossRef]
- Koch, S.; Pasch, S. CultureBERT: Measuring Corporate Culture with Transformer-Based Language. In Proceedings of the 2023 IEEE International Conference on Big Data (BigData), Sorrento, Italy, 15–18 December 2023. [Google Scholar]
- Young, J.; Adams, J.; Baker, N. AI in HR Analytics: Enhancing Decision-Making for Employee Growth. 2025. [Google Scholar]
- Mullens, D.; Shen, S. 2ACT: AI-Accentuated Career Transitions via Skill Bridges. arXiv 2025, arXiv:2505.07914. [Google Scholar] [CrossRef]
- Antoniuk, D.; Ivens, B.S.; Kolyada, O. How is artificial intelligence changing hr? Adaptive management for the new environment. Balt. J. Econ. Stud. 2025, 11, 13–26. [Google Scholar] [CrossRef]
- Do, H.; Chu, L.X.; Shipton, H. How and when AI-driven HRM promotes employee resilience and adaptive performance: A self-determination theory. J. Bus. Res. 2025, 192, 115279. [Google Scholar] [CrossRef]
- Yanamala, K.K.R. Transparency, Privacy, and Accountability in AI-Enhanced HR Processes. J. Adv. Comput. Syst. 2023, 3, 10–18. [Google Scholar] [CrossRef]
The Model | The Prediction |
---|---|
Linear regression | |
Ordinary least squares regression | 0.426 |
Robust regression | 0.423 |
Ridge regression | 0.426 |
Non-linear regression | |
General Gaussian | 0.426 |
General Poisson | 0.426 |
Gamma regression | 0.426 |
Regression tree | 0.500 |
SVM | 0.855 |
Random forest | 0.853 |
Neural network | 0.861 |
Studies (Author (Year) [Ref.]) | Technique | Task | Results |
---|---|---|---|
Andr’e Artelt, et al., 2023 [25] | Random forest, XGBoost | Employee attrition prediction and the generation of counterfactual explanations | An approximate 40% increase in salary would significantly reduce employees’ intention to leave. Reducing the time since the last promotion by about 5 years has been associated with a lower probability of departure. A combination of a 20% salary increase and an approximately 50% increase in job satisfaction would decrease the likelihood of employees leaving. |
Jia Yuan et al., 2022 [45] | Linear regression, non-linear regression, regression tree, SVM (support vector machine), random forest, neural network | Performance prediction, employee attrition | SVM achieved an AUC of 0.855. Random forest achieved an AUC of 0.853. Neural network achieved an AUC of 0.861. Linear regression achieved an AUC of 0.426. Non-linear regression achieved an AUC of 0.426. Regression tree achieved an AUC of 0.500. |
Narayana Darapanen, et al., 2022 [42] | Random forest, XGBoost, support vector machine (SVM), logistic regression and a model ensemble | Employee attrition risk prediction | XGBoost had the best performance with an AUC of 0.86. Ensemble average achieved an AUC of 0.85. Random forest achieved an AUC of 0.81. Logistic regression reached an AUC of 0.83. SVM achieved an AUC of 0.72. |
Bobbinpreet Kaurm et al., 2022 [50] | Decision tree (DT), Boosted tree ensemble, K-nearest neighbor (KNN), support vector machine (SVM), Naive Bayes (NB) | employee retention prediction | Decision tree (DT): 98.0% accuracy. Boosted tree ensemble: 97.5% accuracy. K-nearest neighbor (KNN): 93.6% accuracy. Naive Bayes (NB): 90.3% accuracy. Support vector machine (SVM): 78.1% accuracy. |
Yong Shi, et al., 2022 [49] | Support vector machine (SVM), Naive Bayes, decision tree, neural network | Employee turnover prediction | Decision tree: accuracy 97.03%, precision 97.54%, recall 98.61%, F1 score 98.07%. Neural network: accuracy 93.88%, precision 95.58%, recall 96.47%, F1 score 96.02% Naive Bayes: accuracy 78.94%, precision 90.43%, recall 80.97%, F1 score 85.44%. Support vector machine (SVM): accuracy 78.11%, precision 79.75%, recall 95.61%, F1 score 86.92%. |
Vengai Musanga et al., 2022 [46] | Logistic regression (LR), random forest (RF), gradient boosting machine (GBM), decision tree (DT) and K-nearest neighbors (KNN). | Employee churn prediction | With feature selection (Pearson correlation): RF achieved the highest accuracy, 91.76%. Without feature selection: GBM had the highest accuracy, 87.53%. ROC AUC: RF and GBM showed the best results, with large areas under the ROC curves, indicating a high level of separability. |
Heng Zhang, et al., 2018 [54] | Logistic regression, gradient boosted decision tees (GBDT) | Employee turnover prediction | Logistic model accuracy: 87.2% Gradient boosted decision trees accuracy: 89.32%. |
Shobhanam Krishna, et al., 2022 [41] | Random forest classifier, AdaBoost classifier | Employee attrition prediction | Random forest classifier accuracy 100%. AdaBoost classifier accuracy 84.95%. |
Isha Tewari, et al., 2020 [24] | Random forest (RF), gradient boosting machines (GBM), decision trees (DT), support vector machines (SVM), and k-nearest neighbors (kNN) | Predicting employee attrition | RF achieved an accuracy of 91.76% with feature selection. GBM showed robust performance with an accuracy of 87.53% without feature selection. |
Said Achchab, et al., 2021 [21] | Machine learning algorithms: primarily used for predictive analytics in recruitment, retention, and performance assessments. KNN, logistic regression, SVM Natural language processing (NLP): For analyzing text data, such as resumes, cover letters, and employee feedback. Robotic process automation (RPA): for automating repetitive tasks in HR operations, such as data entry, scheduling, and initial candidate screening. | Predicting employee turnover and enhancing HR efficiency | Increased efficiency: AI reduces manual work in HR, leading to time and cost savings. Enhanced decision-making: predictive analytics enables HR departments to identify high-risk turnover employees and make proactive interventions. KNN 95%, logistic regression 95%, SVM 5% Improved recruitment accuracy: AI algorithms, such as NLP-based models, enhance candidate screening and matching processes. |
Abdullah Abonamah, et al., 2022 [38] | The system architecture comprises three main components: dashboard service: provides a visual interface to HR professionals, displaying key metrics and predictions. ML API services: uses RESTful APIs for delivering machine learning models, focusing on attrition prediction. Database service: stores HR data using PostgreSQL, integrated via an open database connectivity (ODBC) API to support various DBMS. The ML pipeline uses logistic regression as the main predictive model, and AdaBoost, gradient boosting, support vector machine, linear discriminant analysis | The primary task is predicting employee attrition. This involves: Data preparation: handling class imbalance in the attrition dataset and selecting relevant features based on statistical tests. Model training and calibration: logistic regression was chosen as the best model, calibrated with Platt scaling to interpret predictions as probabilities, especially for cases with high stakes like employee attrition. | The calibrated logistic regression model provided better-balanced predictions by reducing false positives. The accuracy is 0.854—logistic regression, 0.822—AdaBoost, 0.847—gradient boosting, 0.849—support vector machine, 0.849—linear discriminant analysis. |
Shweta Pandey, et al., 2022 [52] | Modeling utilizes algorithms like random forest, Naïve Bayes, K-nearest neighbors (KNN), and logistic regression. | Using employee characteristics to predict which employees are suitable for promotion, assisting HR in timely and accurate decision-making. | The random forest model was found to have the highest accuracy for prediction tasks. Random forests—99.6%. KNN—89%. Naïve bayes—73.4%. Logistic regression—60%. |
Priyanka Sadana, et al., 2021 [39] | Various models—logistic regression, random forest, decision tree, and gradient boosting—were trained and evaluated to determine the best predictor. | The primary task was to predict employee attrition and identify factors that most influence an employee’s decision to stay or leave. | The random forest model outperformed others, achieving high recall (93% post-tuning), indicating reliable predictions for attrition. |
Costa, R., et al., 2022 [51] | Machine learning models used are decision tree, AdaBoost, and support vector machine (SVM). | The main task of the model is to predict whether an employee is likely to leave the organization. This allows HR teams to proactively address potential turnover by providing targeted support, incentives, or role adjustments. | The decision tree model achieved the highest accuracy (AUC = 0.78), followed by AdaBoost (AUC = 0.74). SVM was not suitable for this problem, with an AUC of 0.50, indicating it struggled to differentiate between retention and turnover. |
Apurva BM HR Analytics, et al., 2020 [43] | Six machine learning models (logistic regression, decision tree, K-nearest neighbors, Naïve Bayes, support vector machine, and XGBoost) were tested. The dataset was split 70:30 for training and testing, and the models were evaluated using metrics like accuracy, F1-score, and ROC-AUC. | Identify employees likely to leave based on factors like happiness index, salary, and work environment. | XGBoost outperformed other models with an accuracy of 96%, AUC of 0.95, and F1-score of 0.96, making it the preferred model for attrition prediction. AUC obtained for models: logistic regression—0.71 decision tree—0.71 K-nearest neighbors—0.56 Naïve Bayes—0. 67 support vector machine—0.53 XGBoost—0.95 |
Sarah S. Alduayj, et al., 2018 [53] | Support vector machine (SVM), random forest, and K-nearest neighbors (KNN) | Predict employee attrition by analyzing factors that contribute to the likelihood of an employee leaving. | Imbalanced dataset: quadratic SVM performed best with an F1 score of 0.503, but overall, the imbalanced dataset yielded low F1 scores across all models. ADASYN balanced dataset: results significantly improved, with cubic SVM, Gaussian SVM, random forest, and KNN (K = 3) achieving F1 scores between 0.91 and 0.93. This demonstrated the effectiveness of the ADASYN oversampling technique. Feature selection: using feature ranking, top features (like overtime, total working years, and job level) were identified. Random forest achieved an F1 score of 0.909 using the top 12 features, reducing model complexity without sacrificing performance. Undersampled dataset: performance was lower than ADASYN, with Gaussian SVM achieving the best F1 score of 0.738. This indicated that undersampling might lead to information loss, affecting model performance. |
Amine Habous, et al., 2021 [47] | Several classification algorithms were used, including: Gaussian Naïve Bayes Bernoulli Naïve Bayes Multinomial Naïve Bayes Decision tree Random forest Logistic regression | The task was to predict employee attrition and minimize false negatives (i.e., accurately identify employees who are at risk of leaving). | The task was to predict employee attrition and minimize false negatives (i.e., accurately identify employees who are at risk of leaving). |
Raj Chakraborty, et al., 2021 [16] | algorithms used: random forest, logistic regression, Naïve Bayes, gradient boosting, SVM, and K-nearest neighbors. | Understanding the key factors driving employee turnover and developing efficient predictive models. | The random forest model delivered the best results with an accuracy of 90.2%, while Naïve Bayes had the lowest accuracy at 80.7%. |
Preethi Keerthi DSouza, 2023 [58] | Bayesian network | Employee attrition analysis and prediction | Percent correct classification for attrition = 89% |
Model | Relation Type | Sensitivity to Imbalance | AUC |
---|---|---|---|
Logistic regression | Linear | High | Lower |
Random forest | Non-linear | Medium | Good |
XGBoost | Non-linear | Low | Best |
Model | AUC | Accuracy | Comments |
---|---|---|---|
Logistic regression | 0.70–0.85 | 79–82% | Performs well on linear and well-scaled datasets; less effective with complex, nonlinear data. |
Random forest | 0.80–0.91 | 78–89% | Stable and interpretable; prone to overfitting without regularization. |
Support vector machine (SVM) | 0.65–0.88 | 68–85% | Performance varies significantly across studies. |
ExtraTrees | 0.83–0.95 | 80–92% | Improved generalization and robustness to noise; suitable for HR datasets with mixed features. |
XGBoost | 0.96–0.99 | 85–94% | High performance in imbalanced and nonlinear contexts; strong regularization reduces overfitting |
Neural network | 0.84–0.92 | 81–90% | Good at capturing complex patterns; requires more training time and larger datasets. |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Căvescu, A.M.; Popescu, N. Predictive Analytics in Human Resources Management: Evaluating AIHR’s Role in Talent Retention. AppliedMath 2025, 5, 99. https://doi.org/10.3390/appliedmath5030099
Căvescu AM, Popescu N. Predictive Analytics in Human Resources Management: Evaluating AIHR’s Role in Talent Retention. AppliedMath. 2025; 5(3):99. https://doi.org/10.3390/appliedmath5030099
Chicago/Turabian StyleCăvescu, Ana Maria, and Nirvana Popescu. 2025. "Predictive Analytics in Human Resources Management: Evaluating AIHR’s Role in Talent Retention" AppliedMath 5, no. 3: 99. https://doi.org/10.3390/appliedmath5030099
APA StyleCăvescu, A. M., & Popescu, N. (2025). Predictive Analytics in Human Resources Management: Evaluating AIHR’s Role in Talent Retention. AppliedMath, 5(3), 99. https://doi.org/10.3390/appliedmath5030099