Research Article
Price Estimation of Used Cars Using Machine Learning Algorithms
@INPROCEEDINGS{10.1007/978-3-031-28975-0_3, author={B. Valarmathi and N. Srinivasa Gupta and K. Santhi and T. Chellatamilan and A. Kavitha and Armaan Raahil and N. Padmavathy}, title={Price Estimation of Used Cars Using Machine Learning Algorithms}, proceedings={Cognitive Computing and Cyber Physical Systems. Third EAI International Conference, IC4S 2022, Virtual Event, November 26-27, 2022, Proceedings}, proceedings_a={IC4S}, year={2023}, month={3}, keywords={CatBoost regression Support Vector regression Random Forest regression Machine learning}, doi={10.1007/978-3-031-28975-0_3} }- B. Valarmathi
N. Srinivasa Gupta
K. Santhi
T. Chellatamilan
A. Kavitha
Armaan Raahil
N. Padmavathy
Year: 2023
Price Estimation of Used Cars Using Machine Learning Algorithms
IC4S
Springer
DOI: 10.1007/978-3-031-28975-0_3
Abstract
In this study, machine learning (ML) techniques are employed to predict used car prices. Several features are used to calculate the price of used cars, but in this paper, we find efficient ways to find the most precise car prices. Despite the fact that there are websites offering this service, they could not employ the most precise prediction system. It is also possible to predict a used car's true market value using a variety of models and techniques. It's important to understand their genuine market value before buying or selling. Both buyers and sellers will be benefitted from these accurate predictions. Support Vector regression, Random Forest regression, and CatBoost regression techniques are used in the proposed system. In the existing method [13], mean absolute error for decision tree regression was 0.6711, which was the least among other algorithms like Linear regression, Lasso regression, Ridge regression, Bayesian Ridge regression, and etc., they used. In the proposed system, mean absolute error (MAE) for Support Vector regression, CatBoost regression and Random Forest regression techniques are 0.1459, 0.1371 and 0.1284 respectively. The prices of second hand/used cars are predicted using the CatBoost regression, Support Vector regression, and Random Forest regression techniques. The accuracy of these algorithms are 86.28%, 85.40% and 87.16%. Among these three algorithms, Random Forest regression gives the least MAE of 0.1317 and the highest accuracy of 87.16%.
