Prediction of Used Cars Prices Using Machine Learning and Large Language Models
Research Digest on Engineering Management and Social Innovations · 10.04.2026 · 2 (4)
DOI: 10.46647/icetetas170 ↗Kurz gefasst
Fachartikel in der Zeitschrift Research Digest on Engineering Management and Social Innovations, online erschienen am 10.04.2026. Verfasst von B. Pavani und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Used cars sales business has seen an explosion in the volumes in the recent years largely driven by the digital evolution & the numerous online marketplaces that have come up in India such as cars24, OLX, cars, etc. and so on. The online marketplaces have greatly streamlined the whole process of buying and selling a car. But still there lies lots of scope of improvement to price integrity of the car across various stakeholders. Price evaluation of a used car is a highly challenging task that should take into account various parameters such as the age of the car, its mileage, manufacturer, type of engine and transmission. Usually, the assessment of the value of a used car is highly subjective. A data driven approach may therefore be highly beneficial for the assessment of the value of a second hand car in order to provide the seller with an objective value. In this paper, a novel, data science method for estimation of a used car value is introduced. The approach is based on the estimation of a car value through the utilisation of a large data set of car values, obtained from real-world used cars listings over the past years. Data Preprocessing, Feature Engineering & EDA The proposed system will utilise the pre-processed data, engineered features and results obtained from EDA to choose the relevant features that impact the final price. It will then train a variety of regression models including Linear Regression, Decision Tree Regressor, Random Forest Regressor and Gradient Boosting Regressor and choose the one that yields the least: Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) and maximises the coefficient of determination (R²). Real-world experiments have demonstrated that many popular machine learning methods like Random Forest and Gradient Boosting have surpassed linear models in terms of accuracy in pricing used cars and other products. The proposed method is not only highly accurate but also very reliable, scalable and interpretable in real-world used car pricing applications.
Zitieren
APA
Pavani, B., & Rani, M. B. (2026). Prediction of Used Cars Prices Using Machine Learning and Large Language Models. Research Digest on Engineering Management and Social Innovations, 2(4). https://doi.org/10.46647/icetetas170
Vancouver
Pavani B, Rani MB. Prediction of Used Cars Prices Using Machine Learning and Large Language Models. Research Digest on Engineering Management and Social Innovations. 2026;2(4). doi:10.46647/icetetas170
BibTeX
@article{Pavani2026,
author = {Pavani, B. and Rani, M. Beulah},
title = {Prediction of Used Cars Prices Using Machine Learning and Large Language Models},
journal = {Research Digest on Engineering Management and Social Innovations},
year = {2026},
volume = {2},
number = {4},
doi = {10.46647/icetetas170}
}
EUVIMED