Abstract :

In Todays Scenario, The Workforce In IT Sector Has Been Experiencing Uncertainty Because Of The Changing Need Of The Business, Organisational Restructuring And Changing Project Need; Hence, It Is An Important Area Of Human Resource Analytics To Predict Whether An Employee Will Be Laid Off Or Not. Knowing Which Employes Are At Risk Of Layoff Helps Organisations Better Plan Their Workforce, Retain Employes And Make Strategic Decisions. The Methodology Developed Is Based On An Employe Layoff Data Set Consisting Of Demographic, Organisational, Performance, Attendance, Income, Project And Work Experience Features. The Steps Taken In The Preparation Pipeline Include Feature Engineering, Label Encoding, Feature Normalisation Using StandardScaler, And Stratified Data Partitioning To Improve Data Quality For Learning And Balancing The Classes With SMOTE. We Comparatively Analyse CatBoost, LightGBM, Stacking Classifier, BiLSTM, Hybrid (BiLSTMCatBoost-LightGBM) And Hybrid (Stacking-BiLSTM) Models With The Help Of Accuracy, Precision, Recall, F1-score And ROCAUC. Based On The Experimental Results, The Hybrid (Stacking– BiLSTM) Model Has The Best Results In Classification With Accuracy Of 96.4% And ROC-AUC Of 0.994. But The One With High Deploying Performance, High Predictive Performance And Low Computational Cost Is Called The Stacking Classifier And Is Selected For The Deployment. To Improve Transparency And User Trust, XAI Methods Such As LIME, SHAP, PDP And ICE Are Applied To Probe The Model Predictions And The Created Web App Using Flask Allows For Layoff Risk Prediction, Confidence Estimation, And Interactive Explanation Visualisations, Aiding Informed Human Resource Decision Making Processes. “Keywords— Employee Layoff Risk Prediction, Human Resource Analytics, Temporal Feature Fusion, Hybrid Machine Learning, Explainable Artificial Intelligence (XAI), Workforce Management.”

Published:

10-8-2026

Issue:

Vol. 26 No. 8 (2026)

Page Nos:

825 - 830

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite