Machine Learning for Used Car Price Prediction in the Egyptian Market

The used car market in Egypt represents a significant sector of the automotive industry, where accurate pricing and market analysis are crucial for both buyers and sellers to make informed decisions. However, this market faces unique challenges, including price volatility, information asymmetry, and the absence of standardized pricing mechanisms, making it difficult for stakeholders to determine fair market values. Traditional regression techniques yielded limited performance with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.81. This paper addresses these limitations through multiple approaches. First, developed an extensive dataset by implementing continuous web scraping mechanisms across major Egyptian online marketplaces, ensuring broad coverage of all available brands and models with real-time updates. Our methodology incorporates thorough data preprocessing to standardize and clean the collected information, followed by the implementation of diverse prediction models utilizing both classical statistical methods and advanced machine learning techniques. Experimental results demonstrate significant improvements over existing approaches, achieving prediction accuracy scores 0.92 using CatBoost algorithm, marking a substantial advancement over previous egyptian market studies in the literature.