From data to sustainability: Using explainable AI to promote electric vehicle development and understand consumer preferences

IF 2.1 3区经济学 Q2 DEVELOPMENT STUDIES

Development Policy Review Pub Date : 2026-01-30 DOI:10.1111/dpr.70059

Xianglei Zhu, Jing Yang, Fan Zhang

{"title":"From data to sustainability: Using explainable AI to promote electric vehicle development and understand consumer preferences","authors":"Xianglei Zhu, Jing Yang, Fan Zhang","doi":"10.1111/dpr.70059","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> Motivation</h3>\n \n <p>The global need to accelerate the development and consumption of electric vehicles (EVs) as a sustainable alternative to traditional transport.</p>\n </section>\n \n <section>\n \n <h3> Purpose</h3>\n \n <p>To identify potential consumers willing to switch to EVs and understand the underlying drivers of their behavior by utilizing Explainable Artificial Intelligence (EAI).</p>\n </section>\n \n <section>\n \n <h3> Approach and methods</h3>\n \n <p>The study analysed data from 1,964 users. Technical characteristics (Class A) were evaluated using Probit regression and one-way ANOVA, while sociodemographic characteristics (Class B) were assessed via logistic regression. A user mining model was developed to predict behaviour, with Shapley Additive exPlanations (SHAP) used to determine the relative contribution of each predictor. Additionally, simulated enhancements in product indicators were modeled to identify “persuadable” potential buyers.</p>\n </section>\n \n <section>\n \n <h3> Findings</h3>\n \n <p>Purchasing decisions are significantly influenced by a combination of sociodemographic and technical factors. SHAP analysis successfully quantified the impact of these predictors, and potential user modeling confirmed that improvements in specific technical features could effectively convert reluctant consumers into EV adopters.</p>\n </section>\n \n <section>\n \n <h3> Policy implications</h3>\n \n <p>To drive EV adoption, developers and policy-makers should prioritize targeted technical enhancements. Marketing and product development strategies should focus on those specific technical indicators that the model shows have the highest potential to shift consumer behaviour towards adoption.</p>\n </section>\n </div>","PeriodicalId":51478,"journal":{"name":"Development Policy Review","volume":"44 2","pages":""},"PeriodicalIF":2.1000,"publicationDate":"2026-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Development Policy Review","FirstCategoryId":"96","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1111/dpr.70059","RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"DEVELOPMENT STUDIES","Score":null,"Total":0}

引用次数: 0

Abstract

Motivation

The global need to accelerate the development and consumption of electric vehicles (EVs) as a sustainable alternative to traditional transport.

Purpose

To identify potential consumers willing to switch to EVs and understand the underlying drivers of their behavior by utilizing Explainable Artificial Intelligence (EAI).

Approach and methods

The study analysed data from 1,964 users. Technical characteristics (Class A) were evaluated using Probit regression and one-way ANOVA, while sociodemographic characteristics (Class B) were assessed via logistic regression. A user mining model was developed to predict behaviour, with Shapley Additive exPlanations (SHAP) used to determine the relative contribution of each predictor. Additionally, simulated enhancements in product indicators were modeled to identify “persuadable” potential buyers.

Findings

Purchasing decisions are significantly influenced by a combination of sociodemographic and technical factors. SHAP analysis successfully quantified the impact of these predictors, and potential user modeling confirmed that improvements in specific technical features could effectively convert reluctant consumers into EV adopters.

Policy implications

To drive EV adoption, developers and policy-makers should prioritize targeted technical enhancements. Marketing and product development strategies should focus on those specific technical indicators that the model shows have the highest potential to shift consumer behaviour towards adoption.

查看原文本刊更多论文

从数据到可持续性：利用可解释的人工智能促进电动汽车发展，了解消费者偏好

全球需要加速电动汽车（ev）的发展和消费，作为传统交通工具的可持续替代品。目的利用可解释人工智能（EAI）识别愿意转向电动汽车的潜在消费者，并了解其行为的潜在驱动因素。研究分析了1964名用户的数据。技术特征（A类）使用Probit回归和单因素方差分析进行评估，而社会人口特征（B类）通过逻辑回归进行评估。开发了一个用户挖掘模型来预测行为，使用Shapley加性解释（SHAP）来确定每个预测器的相对贡献。此外，模拟产品指标的增强，以确定“可说服”的潜在买家。调查结果：消费者的购买决策受到社会人口因素和技术因素的共同影响。SHAP分析成功地量化了这些预测因素的影响，潜在用户建模证实了特定技术特征的改进可以有效地将不情愿的消费者转变为电动汽车的采用者。为了推动电动汽车的普及，开发人员和政策制定者应该优先考虑有针对性的技术改进。营销和产品开发战略应侧重于模型显示最有可能使消费者行为转向采用的那些具体技术指标。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Development Policy Review DEVELOPMENT STUDIES-

CiteScore

3.50

自引率

5.90%

发文量

期刊介绍： Development Policy Review is the refereed journal that makes the crucial links between research and policy in international development. Edited by staff of the Overseas Development Institute, the London-based think-tank on international development and humanitarian issues, it publishes single articles and theme issues on topics at the forefront of current development policy debate. Coverage includes the latest thinking and research on poverty-reduction strategies, inequality and social exclusion, property rights and sustainable livelihoods, globalisation in trade and finance, and the reform of global governance. Informed, rigorous, multi-disciplinary and up-to-the-minute, DPR is an indispensable tool for development researchers and practitioners alike.