Predicting the Regional Adoption of Electric Vehicle (EV) With Comprehensive Models
Jianmin Jia, Baiying Shi, Fa Che, Hui Zhang
Abstract
Adoption of electric vehicles (EVs) has been regarded as one of the most important strategies to address the issues of energy dependence and greenhouse effect. Empirical reviews demonstrate that wide acceptance of EV is still difficult to achieve. This research proposes to investigate the factors that might trigger the wide usage of EVs to support the energy policy. The real-world owners of EV were extracted from the 2017 National Household Travel Survey (NHTS), which provides large-scale individual characteristics. NHTS dataset was processed to establish the comprehensive estimation model for EV adoption with considering vehicle, personal and household factors. Besides the commonly social-economic factors, the gasoline price and car sharing program were found to be significant for EV adoption. Additionally, since the EV owners are only 1.29% of all vehicle owners, this article introduced the imbalanced dataset technique, which was seldom considered in existing researches. Subsequently, several machine learning methods were utilized to build the prediction model, and the model performance analysis indicates the Decision Tree (DT) model outperforms other models. A regional EV penetration map was also generated for the U.S. to validate the proposed approach. Implications for further research, transport policy and EV market are discussed.