<![CDATA[Although Single Sign On (SSO) simplifies access to Human Capital Management Information Systems (HCMIS), Employee Self Service (ESS) adoption remains a challenge in many organizations. This study aims to develop a machine learning framework for predicting ESS adoption and SSO usage and identifying the factors that influence employee digital engagement. A quantitative experimental approach was applied to a simulated dataset of 3,000 employee records containing demographic, behavioral, and system utilization variables. Data preprocessing included missing value imputation, categorical encoding, normalization, an 80:20 training and testing split, and SMOTE to address class imbalance. Eight classification models were evaluated using Accuracy, Precision, Recall, F1 score, and AUC ROC. The results show that SMOTEBoost achieved the best performance, with an Accuracy of 0.924, F1 score of 0.908, and AUC ROC of 0.956, followed by RUSBoost with an F1 score of 0.901 and AUC ROC of 0.951, and Balanced Random Forest with an F1 score of 0.897 and AUC ROC of 0.947. ESS Login Frequency, SSO Usage Frequency, and Training Participation were the strongest predictors of adoption. The proposed framework complements technology adoption perspectives such as TAM and UTAUT by moving beyond survey based explanations of perceptions and intentions toward predictive analysis using observable system utilization and behavioral data. This approach enables organizations to identify employees with lower adoption potential earlier and provide more targeted managerial interventions.]]>
