Deciphering the Augmentation of Classification Models in Predicting Employee Attrition

Raj, Aditya (2022) Deciphering the Augmentation of Classification Models in Predicting Employee Attrition. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration manual) Download (3MB) | Preview

Abstract

Workforce administration is at the cusp of renaissance due to the surge in attrition rates globally. Every organization is driven by its employees and their ability to meet their objectives and it is strategically imperative for employers to retain talented and highly skilled professionals. It is critical to monitor employee demographics and contrive a plan to identify potential conflicts in organizational setting. The juxtaposition of attrition and retention align on the analysis of common factors. With the advent of machine learning in the field of HR Analytics, employee attrition prediction has been fulfilled through implementation of classification techniques. The research aims to extend the use of Random Forest classifier and Light Gradient Boosting Machine (LightGBM) classifier to predict employee attrition on an class imbalanced artificially simulated HR Employee dataset using oversampling methods such as SMOTE and Random Oversampling. The classification methods are integrated with Grid Search Cross Validation for Best Hyperparameters, Recursive Feature Elimination (RFE, advanced method for feature selection) and Manual Hyperparameter tuning. The models are evaluated based on F1 Score, Accuracy and ROC AUC score. It is concluded that Random Forest classifier with manual tuning of Hyperparameter performed better than the Grid Search Cross Validation and Recursive Feature Elimination approaches with a F1 score of 62.18% , Accuracy of 84.69% and ROC AUC score of 84.28% whereas manually hyperparameter tuned LightGBM classifier exhibited incredible performance with a F1 score of 65.91%, Accuracy of 89.8% and ROC AUC score of 78.42%.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Class Imbalance; Classification; Employee Attrition; Hyperparameter Tuning; Recursive Feature Elimination; SMOTE
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management > Human Resource Management > Employee Retention Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning H Social Sciences > HF Commerce > Personnel Management
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	01 Mar 2023 12:06
Last Modified:	01 Mar 2023 17:31
URI:	https://norma.ncirl.ie/id/eprint/6267

Actions (login required)

View Item