Parametric and non-parametric learning models for time series: case study of vehicle sales based on exogenous variables in Brazil

Abstract Demand forecasting is a strategic necessity for vehicle manufacturers, directly influencing production schedules, operations management, and resource allocation. As markets become increasingly volatile, advanced mathematical models, particularly those leveraging machine learning techniques, have become indispensable. This study presents a comprehensive comparative analysis of parametric and non-parametric models for time series forecasting of vehicle demand in Brazil. By integrating key economic indicators, such as GDP growth, unemployment rates, and interest rates, we explore how these models can more accurately predict market demand and adapt to external economic variables. Through extensive computational simulations, we evaluate the accuracy, robustness, and practical applicability of these models. Our findings provide critical insights into the strengths and limitations of each modeling approach, offering valuable guidance for improving demand forecasting strategies. This research not only affirms the efficacy of learning models in enhancing demand planning but also lays a robust groundwork for future theoretical and practical advancements across various economic sectors.