Comparative Analysis of XGBoost and Random Forest for Used Car Price Prediction

Rapid technological developments have transformed the automotive industry, especially in increasing the availability and demand for used cars. Online platforms like OLX Autos are becoming increasingly popular, especially during the pandemic, as they facilitate transactions between buyers and sellers. However, pricing used cars remains a challenge. Therefore, machine learning can be one of the solutions that can be done, especially the Extreme Gradient Boosting (XGBoost) and Random Forest algorithms, to predict the price of used cars. In our works, the datasets is crawled and scraped from OLX Indonesia used car platform. The prediction evaluation results showed that the Random Forest model yielded excellent performance with hyperparameter configuration n estimators = 100, random state = 0, min samples leaf = 1, max features = 0.3, n jobs = -1, and oob score = True achieved a R-squared value of 0.9485 with a test time of 120 ms. The model is also used in used car price prediction website for different car types.