Price Prediction and Analysis of Price Influencing Factors for Second-hand Car Sales in AutoTrader Based on XGBoost Algorithm

Abstract

The global used car market continues to expand, reaching a scale of 1.6 trillion US dollars in 2023. In 2024, China's transaction volume reached 19.61 million units, setting a new high. However, information asymmetry, sharp price fluctuations, and subjective assessment severely constrain market efficiency. To solve the pricing problem, this study, based on a large amount of data from the AutoTrader platform in the UK, builds an XGBoost high-precision price prediction model, integrates multiple vehicle attributes and market characteristics, and achieves low-error residual value estimation. At the same time, random forest feature analysis is used to quantify the contribution of key factors, revealing the hierarchical influence structure, providing intelligent and data-driven pricing decision support for all parties involved in the transaction, and promoting the market to transform towards transparency and efficiency.