IMPROVED MACHINE LEARNING MODEL FOR VEHICLE PRICE PREDICTION IN THE NIGERIAN ECONOMY

Abstract

The Nigerian used car market is characterized by significant price variability, lack of transparency, and inconsistent valuation mechanisms, posing challenges to both buyers and sellers. This research aimed to develop a robust, data-driven predictive model tailored to the specific dynamics of the Nigerian automotive ecosystem using machine learning algorithms. The study employed a comprehensive dataset of used vehicles listed in Nigeria, incorporating features such as make, model, year, mileage, engine size, fuel type, transmission, condition, and location. Extensive data preprocessing, exploratory analysis, and feature engineering were conducted to uncover the most influential variables affecting vehicle prices. Six machine learning models—Linear Regression, Ridge Regression, Lasso Regression, Decision Tree Regressor, Support Vector Regression, and Random Forest Regressor were trained and evaluated using performance metrics including Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared (R²) Score. The Random Forest Regressor outperformed other models, achieving the highest prediction accuracy with an R² score of 0.91 and the lowest RMSE, making it the most suitable algorithm for this context when compared with previous work done by Pudaruth (2023) and Breen et al. (2024). The study identified vehicle year, mileage, brand, engine size, and geographic location as key determinants of price. The resulting model provides a practical framework for real-time price prediction and can be integrated into digital platforms for use by dealerships, private sellers, and online marketplaces. This research contributes to local automotive market intelligence, promotes pricing transparency, and underscores the transformative potential of machine learning in emerging economies.