This project aims to predict vehicle pricing based on various attributes related to design, performance, market conditions, and temporal factors. It leverages machine learning techniques to provide accurate and actionable insights into automobile pricing.
-
Hyperparameter tuning reduced overall prediction error (MSE/RMSE improvements).
-
R² improved from 95.6% → 96.5%, showing stronger explanatory power
-
Slight MAE increase indicates tuning improved variance reduction more than absolute error minimization.
-
Overall, the tuned Decision Tree is more robust and generalizable.
Design & Performance: Engine Fuel Type, Engine HP, Engine Cylinders, Transmission Type, Driven Wheels, Number of Doors, Vehicle Size, Vehicle Style.
Market & Timing: Make, Model, Year, Market Category, Highway MPG, City MPG.