Best Price Prediction for Pre-Owned Cars using Machine Learning Techniques

In the used-car market that is so competitive, accurate price estimation is critical for both buyers and sellers. Traditional valuation methods struggle with complex relationships between car attributes, market conditions, and economic factors. This study proposes a hybrid stacking model that integrates machine learning algorithms like gradient boosting, random forests, and linear regression to enhance prediction accuracy and robustness. Using a stacked ensemble approach, the model combines the strengths of individual algorithms to improve price estimation and prevent outliers from affecting the model. With a comprehensive dataset that includes car-related features such as mileage, age, brand, and type, feature engineering and predictor significance testing are applied to identify driving factors leading to price fluctuation in the research. The research confirms that the hybrid model makes more precise predictions and outlier detections than individual machine learning models and can therefore serve as a reliable tool for used-car market pricing practice.Keywords: Machine Learning, Price Prediction, Regression Models, Random Forests, Gradient Boosting, Pre-Owned Cars, Market Analysis.