<![CDATA[Customer loyalty in Nigeria’s automotive service sector has become increasingly unstable due to digital competition, pricing inconsistencies, and evolving satisfaction dynamics. Traditional models often overlook the nonlinear relationships shaping loyalty behavior. Most prior research uses linear or descriptive approaches, limiting predictive accuracy and failing to capture behavioral heterogeneity in satisfaction–cost interactions. This constrains proactive customer retention strategies. The study aims to segment customer behavior and predict loyalty determinants using machine learning algorithms to enhance decision-making in automotive service management. Secondary data were obtained from the records department of Anaval Mechanic Workshop, Awka, spanning January to December 2023. The study employed k-Means clustering for behavioral segmentation and machine learning models, Random Forest, Support Vector Machine, and Extreme Gradient Boosting (XGBoost), for loyalty prediction. The XGBoost model achieved the highest predictive accuracy (97.1%) and AUC (0.985). Customer satisfaction, total cost, and non-mechanic service expenses emerged as the strongest loyalty determinants. Machine learning effectively captured nonlinear satisfaction-cost-dynamics, outperforming traditional models. Integrating predictive analytics and cost-transparency frameworks can strengthen retention policies and inform fair-pricing regulations across Nigeria’s automotive service industry.]]>
Copyrights © 2025
