An Explainable Machine Learning Method for Predicting Electric Vehicle Prices in the Iraqi Market
مجلة العلوم الهندسية و تكنولوجيا المعلومات · 15.06.2026 · 10 (2) · 118-135
DOI: 10.26389/ajsrp.t030326 ↗Kurz gefasst
Fachartikel in der Zeitschrift مجلة العلوم الهندسية و تكنولوجيا المعلومات, erschienen am 15.06.2026. Verfasst von Hassan Kassem und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
The emerging market of electric vehicles (EV) in the emerging economies, including in Iraq, is gradually turning into a significant issue of obscure valuation because the conventional internal combustion engine (ICE)-based model of price formation is not appropriate in determining and describing the special determinants of EV price formation. It is against this backdrop that the present study is an endeavor to introduce an explainable machine learning (ML) model which is a precise predictor of prices of used EVs in Iraq, to be more precise, the capital of Baghdad. The methodological strategy presupposed the systematic gathering of much volume of EV data, including such variables as battery capacity, driving range, and a package of local market variables. A regional database of 1420 electric vehicle advertisements on the Iraq market were gathered and assessed. CatBoost, XGBoost, and Random Forest tree-based regression algorithms were utilized to make predictions, and SHapley Additive exPlanations (SHAP) were added to increase transparency and interpretability. The results of the empirical analysis proved that the CatBoost model showed the best predictive performance, with the R2 equal to 0.94, Mean Absolute Error (MAE) equal to 1,245,000 IQD, and Root Mean Squared Error (RMSE) equal to 2,180,000 IQD. Moreover, the findings demonstrate that EV specific attributes, especially battery capacity and driving range are the most influential price factors, which outweigh traditional variables. The supremacy of these characteristics is supported by SHAP analysis which shows these characteristics are supreme. The best-performing model that was tested was CatBoost and it performed particularly well with high-cardinality categorical predictors. Overall, this explicable ML format can improve price transparency and allow informed decision-making among the consumers and dealerships of Iraq and policy makers, which in its turn, has a positive impact on the development of trust and promotes sustainable EV adoption on a regional level.
Zitieren
APA
Kassem, H., & Bishtawi, T. Y. (2026). An Explainable Machine Learning Method for Predicting Electric Vehicle Prices in the Iraqi Market. مجلة العلوم الهندسية و تكنولوجيا المعلومات, 10(2), 118-135. https://doi.org/10.26389/ajsrp.t030326
Vancouver
Kassem H, Bishtawi TY. An Explainable Machine Learning Method for Predicting Electric Vehicle Prices in the Iraqi Market. مجلة العلوم الهندسية و تكنولوجيا المعلومات. 2026;10(2):118-135. doi:10.26389/ajsrp.t030326
BibTeX
@article{Kassem2026,
author = {Kassem, Hassan and Bishtawi, Tariq Yousef},
title = {An Explainable Machine Learning Method for Predicting Electric Vehicle Prices in the Iraqi Market},
journal = {مجلة العلوم الهندسية و تكنولوجيا المعلومات},
year = {2026},
volume = {10},
number = {2},
pages = {118--135},
doi = {10.26389/ajsrp.t030326}
}
EUVIMED