An Intelligent Machine Learning-Based Decision Support Framework for Cloud-Based Human Resource Management and Employee Performance Prediction

Cloud computing coupled with machine learning in Human Resource Management (HRM) has created a new frontier for predicting workforce analytics and making decisions for the organisation based on data. The aim of this work is to explore the supervised machine learning algorithms for predicting employee attrition in a cloud-based HR decision-support system. HR Analytics Employee Attrition and Performance dataset consists of 1,470 employee records and 35 variables relating to the demographic characteristics, professional experience, compensation, job satisfaction, workplace conditions etc. The methodological framework is predicting employee attrition using three models namely Logistic Regression, Random Forest and Extreme Gradient Boosting (XGBoost). Structured Analytical Procedure (SAP) process, data preprocessing, feature engineering, stratified data partitioning and cross-validation are suggested. The objective of model performance evaluation is to evaluate the accuracy, precision, recall, F1 score and ROC-AUC. The framework also contains explainable artificial intelligence (XAI) that can be used to help explain the results of the predictive analysis and to understand what characteristics an employee may have that might be correlated with a risk of attrition. In an organizational standpoint, the suggested approach demonstrates the potential of predictive analysis to aid in employee retention strategies and workforce management decisions. Overall, the study highlights the need for transparency in AI systems, responsible data management, and the role of human oversight in the HR decision-making process. More evidence is needed to confirm the predictive validity and usefulness of these in operational HRM situations.