An enhanced battery materials database auto-generated using ChemDataExtractor and further classified using the BatteryBERT language model.

An enhanced battery materials database auto-generated using ChemDataExtractor and further classified using the BatteryBERT language model. <br> Two battery material databases are included: <br> (1) a new battery property database v2.0 that contains 210,416 data about battery materials and five of its key properties: capacity, voltage, conductivity, Coulombic efficiency and energy. This database supersedes v1.0 (Huang, S. &amp; Cole, J. M. A database of battery materials autogenerated using ChemDataExtractor. <em>figshare</em> https://doi.org/10.6084/m9.figshare.11888115.v2 (2020)) owing to its: (i) enhancement of data from the last two years (2020 and 2021) as well as from papers belonging to Springer journals; (ii) filtering off of non-relevant data from non-battery papers that has been made possible via our implementation of a binary classification model. <br> (2) A device-classified database that contains 300,622 anode, cathode or electrolyte materials that have been classified using our BatteryBERT language model. <br> Both databases have been provided in JSON, SQL, and CSV format.<br>

本增强型电池材料数据库通过ChemDataExtractor自动生成，并借助BatteryBERT语言模型完成进一步分类。<br>本次构建的数据集包含两类电池材料数据库：<br>(1) 新版电池属性数据库v2.0，该库收录210,416条电池材料数据，涵盖容量、电压、电导率、库仑效率（Coulombic efficiency）与能量五大核心属性。本数据库取代了v1.0版本（Huang, S. & Cole, J. M. A database of battery materials autogenerated using ChemDataExtractor. *figshare* https://doi.org/10.6084/m9.figshare.11888115.v2 (2020)），其升级之处包括：(i) 补充了2020、2021近两年的文献数据以及Springer期刊刊载的相关研究内容；(ii) 通过部署二分类模型，过滤了非电池类文献中的无关数据。<br>(2) 按器件分类的数据库，该库收录300,622条经BatteryBERT语言模型分类的负极、正极或电解质材料数据。<br>两类数据库均提供JSON、SQL与CSV三种格式的资源下载。