Topic classification of electric vehicle consumer experiences with transformer-based deep learning.
Topic classification of electric vehicle consumer experiences with transformer-based deep learning.
复制标题
DOI：
10.1016/j.patter.2020.100195
复制
发表时间：
2021-02-12
期刊：
影响因子：
--
通讯作者：
Asensio OI
中科院分区：
文献类型：
--
作者：
Ha S;Marchetto DJ;Dharur S;Asensio OI
交通运输部门是温室气体（GHG）排放的主要来源，也是全球健康不良影响的驱动因素。越来越多的政府政策推动采用电动汽车（ev）作为减少温室气体排放的解决方案。然而，政府分析师未能在与充电基础设施相关的决策中充分利用消费者数据。这是因为很大一部分EV数据是非结构化文本，这给数据发现带来了挑战。在本文中，我们部署了基于转换器的深度学习的进展，以在具有全国代表性的用户评论样本中发现关注的主题。我们报告的分类准确率超过91% (F1分数为0.83)，优于该领域先前领先的算法。我们描述了这些深度学习模型在公共政策分析和大规模实施中的应用。这种能力可以提高电动汽车充电市场的智能化，预计到2027年，电动汽车充电市场将增长到276亿美元。关于电动汽车充电行为的消费者数据是非结构化的，并且在很大程度上处于休眠状态。我们为变压器模型的自动主题分类提供了概念证明，我们达到了91%的准确率（F1 0.83），优于先前领先的算法。看到谷歌搜索和翻译算法在生产级的使用情况。这些模型对许多领域的文本语境学习产生了重大影响，例如医疗保健、金融、制造业；然而，到目前为止，在电动汽车方面还没有经验上的进展。鉴于能源和交通领域的数字化转型，对关键能源基础设施进行实时分析的机会越来越多。一个巨大的、尚未开发的电动汽车移动数据来源是由移动应用程序用户评论充电站生成的非结构化文本。利用基于变压器的深度学习，我们提出了充电站评论的多标签分类，在某些情况下，其性能超过了人类专家。这为自动发现和实时跟踪电动汽车用户体验铺平了道路，这可以为当地和区域政策提供信息，以应对气候变化。政府分析师和政策制定者未能在电动汽车充电基础设施相关决策中充分利用消费者行为数据。这是因为很大一部分EV数据是非结构化文本，这给数据发现带来了挑战。在本文中，我们部署了基于变压器的深度学习的进展，以发现具有全国代表性的电动汽车用户评论样本中的问题。我们描述了公共政策分析的应用，并找到证据表明人口较少的地区在车站可用性方面可能服务不足。
The transportation sector is a major contributor to greenhouse gas (GHG) emissions and is a driver of adverse health effects globally. Increasingly, government policies have promoted the adoption of electric vehicles (EVs) as a solution to mitigate GHG emissions. However, government analysts have failed to fully utilize consumer data in decisions related to charging infrastructure. This is because a large share of EV data is unstructured text, which presents challenges for data discovery. In this article, we deploy advances in transformer-based deep learning to discover topics of attention in a nationally representative sample of user reviews. We report classification accuracies greater than 91% (F1 scores of 0.83), outperforming previously leading algorithms in this domain. We describe applications of these deep learning models for public policy analysis and large-scale implementation. This capability can boost intelligence for the EV charging market, which is expected to grow to US$27.6 billion by 2027. Consumer data on EV charging behavior are unstructured and remain largely dormant We provide proof of concept for automated topic classification with transformer models We achieve 91% accuracy (F1 0.83), outperforming previously leading algorithms Applications for local and regional policy analysis of EV behavior are described Transformer neural networks have emerged as the preeminent models for natural language processing, seeing production-level use with Google search and translation algorithms. These models have had a major impact on context learning from text in many fields, e.g., health care, finance, manufacturing; however, there have been no empirical advances to date in electric mobility. Given the digital transformations in energy and transportation, there are growing opportunities for real-time analysis of critical energy infrastructure. A large, untapped source of EV mobility data is unstructured text generated by mobile app users reviewing charging stations. Using transformer-based deep learning, we present multilabel classification of charging station reviews with performance exceeding human experts in some cases. This paves the way for automatic discovery and real-time tracking of EV user experiences, which can inform local and regional policies to address climate change. Government analysts and policy makers have failed to fully utilize consumer behavior data in decisions related to EV charging infrastructure. This is because a large share of EV data is unstructured text, which presents challenges for data discovery. In this article, we deploy advances in transformer-based deep learning to discover issues in a nationally representative sample of EV user reviews. We describe applications for public policy analysis and find evidence that less populated areas could be underserved in station availability.
登录
查看更多内容
影响因子：
9
作者：
Egbue, Ona;Long, Suzanna
通讯作者：
Long, Suzanna
DOI：
10.1016/j.trd.2012.09.007
发表时间：
2013-01-01
影响因子：
7.6
作者：
Carley, Sanya;Krause, Rachel M.;Graham, John D.
通讯作者：
Graham, John D.
影响因子：
6.7
作者：
Brueckmann, Gracia;Bernauer, Thomas
通讯作者：
Bernauer, Thomas
影响因子：
56.7
作者：
de Rubens, Gerardo Zarazua;Noel, Lance;Sovacool, Benjamin K.
通讯作者：
Sovacool, Benjamin K.
影响因子：
2.9
作者：
Hidrue, Michael K.;Parsons, George R.;Gardner, Meryl P.
通讯作者：
Gardner, Meryl P.