Absenteeism Prediction Using HR Analytics in the Nigerian Manufacturing Sector: A Predictive Modelling Study

This study develops and validates an absenteeism prediction model using HR analytics approaches in the Nigerian manufacturing sector, addressing the operational and financial costs of unplanned workforce absences in a context where HR analytics adoption remains nascent. The study adopts a quantitative predictive design, drawing on secondary HR data from three manufacturing firms in Lagos, Kano, and Enugu States spanning a 36-month period. The dataset includes 3,200 employee records encompassing attendance patterns, performance ratings, commute distances, department classifications, tenure, and health claim frequencies. Logistic regression and decision tree algorithms are developed and compared for absenteeism prediction accuracy. The HR Analytics Framework and Predictive HR Decision-Making Theory provide the conceptual grounding. Findings reveal that the logistic regression model achieves 78.4 percent predictive accuracy, while the decision tree model achieves 81.2 percent, with the latter demonstrating superior sensitivity in identifying high-risk absenteeism profiles. The strongest predictors of chronic absenteeism include frequency of prior disciplinary actions, health claim history, and daily commute distance exceeding 25 kilometres. Seasonal patterns show elevated absenteeism in the first and third quarters, aligned with school resumption and harvest periods respectively. The study recommends deploying predictive absenteeism dashboards as real-time HR management tools, targeting prevention resources at algorithmically identified high-risk employees, and addressing structural commute challenges through transport provision or flexible scheduling. Keywords: HR Analytics, Absenteeism Prediction, Manufacturing, Predictive Modelling, Nigeria.