Forecasting Future EV Sales: A Comparative Study of Model Performance

Christopher Genzhuo Cui

2024

Abstract

The quick advancement of Electric Cars (EVs) globally has sizable implications for consumers, the surroundings, and the auto business enterprise. This study ambitions to are waiting for destiny EV income tendencies via making use of device getting to know fashions—Linear Regression, Random Forest, and Gradient Boosting—to a complete dataset from Kaggle masking global EV income from 2010 to 2024. The dataset becomes meticulously preprocessed via filtering relevant parameters, coping with missing values, normalization, and one-warm encoding of specific variables consisting of vicinity, mode, and powertrain. Every model changed into education and evaluated the usage of advocate Squared errors (MSE) and the R² to assess predictive overall performance. The consequences advocate that artificial intelligence-based ensemble machine learning models like Random Forest and Gradient Boosting barely outperform Linear Regression, accomplishing R² values of approximately 0.18 as compared to 0.08 for Linear Regression. No matter the modest predictive power, those findings spotlight the complexity of modelling EV earnings developments due to factors like insurance changes, monetary conditions, and technological advancements that might not be absolutely captured in the dataset. The look at underscores the potential of the system gaining knowledge of forecasting marketplace developments whilst emphasizing the need for superior feature engineering and model tuning.

Paper Citation

in Harvard Style

Cui C. (2024). Forecasting Future EV Sales: A Comparative Study of Model Performance. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 75-79. DOI: 10.5220/0013487700004619

in Bibtex Style

@conference{daml24,
author={Christopher Genzhuo Cui},
title={Forecasting Future EV Sales: A Comparative Study of Model Performance},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={75-79},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013487700004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Forecasting Future EV Sales: A Comparative Study of Model Performance
SN - 978-989-758-754-2
AU - Cui C.
PY - 2024
SP - 75
EP - 79
DO - 10.5220/0013487700004619
PB - SciTePress