Employee Attrition: Leveraging Machine Learning Explainability to Identify Key Predictors of Turnover
Loading...
Authors
Nguyen, Dinh Cuong (Ernest)
Lovisolo, Gregory
Dr. Tenney, Dan
Issue Date
2025-04-04
Type
Other
Language
en_US
Keywords
Employee attrition prediction , Machine learning explainability , Human resources analytics
Alternative Title
Abstract
The IBM Employee Attrition dataset is a publicly available, real-world dataset from a global technology company, offering a holistic view of the working environment. The target variable in this dataset is Attrition, which has values ‘Yes’ (indicating the employee left) and ‘No’ (indicating the employee stayed). This dataset was prepared for machine learning and split into two parts: training (80%) and test (20%). High turnover rates incur significant financial costs, and employee attrition can disrupt organizational dynamics, lower morale, and negatively impact overall organizational performance. Traditional machine learning models such as decision trees and random forests can accurately predict employee attrition, but they often fail to provide transparency into the factors driving these predictions. Machine Learning Explainability addresses the limitations of traditional machine learning by augmenting decision-making and providing insights into how machine learning models make their predictions.
Description
UB Rise 2025
School of Engineering