This paper presents a data-driven methodology employing machine learning techniques to assess the impact of electric vehicle adoption policies on carbon dioxide (CO2) concentration levels in air quality. The proposed method integrates both classification and time series approaches to analyse the differences in CO2 concentrations resulting from policy implementation and non-implementation scenarios. Historical data from the Urban Observatory in Newcastle upon Tyne, United Kingdom (2017–2019), is used to train machine learning models, considering assumptions from the UK National Grid's future energy scenarios. Four classification methods and a correlation method are used to identify key features for training the time series models. Long Short-Term Memory and Gated Recurrent Unit are used as time series models to predict monthly CO2 concentrations under the policy and non-policy scenarios. The performance of the classification methods is evaluated using confusion matrix, accuracy, F1, recall, and precision scores, while time series models are assessed using Mean Squared Error, Mean Square Log Error, and Root Mean Squared Error. The results highlight the potential of this datadriven approach for validating policies before implementation.
Estimating the Impact of Electric Vehicle Adoption on CO2 Concentrations Using Classification and Time Series Methods
2023-09-24
1801988 byte
Conference paper
Electronic Resource
English
Estimating potential adoption rate of electric vehicles in urban logistics
Taylor & Francis Verlag | 2024
|Impact of policy incentives on the adoption of electric vehicle in China
Elsevier | 2023
|Dynamic Wireless Power Transfer: Potential Impact on Plug-in Electric Vehicle Adoption
SAE Technical Papers | 2014
|Dynamic Wireless Power Transfer: Potential Impact on Plug-in Electric Vehicle Adoption
British Library Conference Proceedings | 2014
|Expanding Electric Vehicle Adoption in Disadvantaged Communities
Transportation Research Record | 2024
|