Published Paper Details:

The used car market has expanded rapidly over the last decade, driven by rising new car prices, improved vehicle durability, digital marketplaces, and changing consumer preferences. This paper analyses the determinants of the selling price of used cars using insights from existing empirical studies and typical datasets collected from online classified platforms and dealer inventories. Key explanatory variables considered include vehicle age, mileage, brand category, model type, fuel type, transmission, vehicle condition, service history, accident history, and cosmetic attributes such as colour and optional features. Prior research consistently shows that mileage and age are the dominant price drivers, while brand reputation and vehicle class exert substantial additional influence; cosmetic factors such as colour typically have limited but sometimes locally relevant effects. Multiple linear regression and machine learning models, such as random forest regression and neural networks, are commonly used to model the relationship between vehicle characteristics and selling price, achieving reasonably low prediction errors suitable for pricing support in practice. The findings discussed in this paper highlight that: o Price declines non linearly with age and accumulated mileage, with sharper depreciation in early years. o Premium and reliable brands command persistent price premiums, even after controlling for age and mileage. o Good vehicle condition, documented service history, and clean accident records substantially raise achievable selling prices. o Machine learning models can improve prediction accuracy compared with simple linear models but at the cost of interpretability. These insights can help consumers make more informed purchase and sale decisions, and assist dealers and platforms in setting optimal prices, managing inventory, and designing data driven pricing tools.