Performance Evaluation of Hyperparameter-optimized Random Forest Regressor for Used Car Price Prediction

Abstract

This study evaluates the performance of a hyperparameter-optimized random forest algorithm for predicting used car prices. The primary objective is to develop a robust model capable of delivering accurate price predictions. The random forest technique was selected for this study due to its proven effectiveness in handling non-linear regression problems. We utilized a public dataset from Kaggle, sourced via web scraping from Mobil123. The GridSearchCV method was employed to tune the hyperparameters and identify the optimal model configuration. The resulting model demonstrates strong predictive power, explaining over 93% of the price variance (R² score > 0.9349). Furthermore, the model's robustness is confirmed by an average cross-validation score of 0.9418. These results affirm that the optimized random forest model is a highly effective tool for this application. This research has practical implications for the automotive market, providing both buyers and sellers with a data-driven tool for more accurate price valuation