Data-Driven Strategies for Predicting and Preventing Employee Turnover
DOI:
https://doi.org/10.56536/jicet.v4i2.133Keywords:
Attrition prediction, HR Analytics, IBM HR Dataset, Machine Learning algorithms, Modern workforce challenges, Support Vector Machine (SVC), Neural Networks, HR Interventions, Model TrainingAbstract
The dynamics of the modern workforce are diverse and complex, with many challenges. Employee turnover is currently a significant concern for organizations as it negatively impacts various aspects. It leads to increased hiring time and costs, reduces operational efficiency, causes disruptions within teams, and damages the organization's reputation as an employer. Consequently, it also impacts the organization's ability to attract talented individuals from the market. This study uses the dataset of employee turnover created by IBM. The study aims to mark key features that help indicate the signs of attrition and find the best-fit algorithm that can be reliable to proceed with HR interventions in indicated areas. The Study examines 13 different algorithms and evaluates their performance where Support Vector Machine (SVC), Neural Networks & logistic Regression outperform other algorithms with an accuracy of 91.40% and 90%. Hence, these models can be utilized by the Human Resources (HR) department to support their decisions about HR interventions to control attrition on what they have received in terms of relevant information concerning this candidate.
