Transparent AI for car Price Prediction Balancing Accuracy and Interpretability

This study introduces an artificial intelligence system which predicts used car prices in Morocco to overcome the weaknesses of conventional valuation techniques that produce subjective and inconsistent results. The evaluation of multiple machine learning models resulted in CatBoost being chosen as the most effective model for prediction purposes. The final model reached an R2 score of 91.35% after processing a web-scraped dataset from local automotive platforms through extensive preprocessing and hyperparameter tuning. The application of SHapley Additive exPlanations (SHAP) ensured model transparency. The system operates through a web-based interface which provides instant price forecasts together with analytical explanations and automated systems for monitoring data shifts and market adjustments. The study shows that combining advanced prediction capabilities with transparent models and ongoing monitoring systems enhances fairness, accountability, and user trust in AI-based pricing systems for emerging markets like Morocco.