Determinants of Used Car Prices: A Multiple Linear Regression Approach

This study analyzes used car pricing determinants using multiple regression techniques (OLS, Ridge, Lasso) and Random Forest machine learning, finding that manufacturing year, mileage, and transmission type are the strongest price predictors, with the Random Forest model achieving the best performance (R² = 0.85). The research provides practical applications for automated valuation tools, helping consumers identify fair prices, dealers optimize inventory pricing, and financial institutions assess loan collateral, while demonstrating that ensemble methods outperform traditional linear models by capturing nonlinear relationships in vehicle pricing data.