five

LLM-Generated User Friendly HS Code Labels for U.S. Trade in 2024

收藏
DataCite Commons2025-03-24 更新2025-04-15 收录
下载链接:
https://dataverse.harvard.edu/citation?persistentId=doi:10.7910/DVN/UVM9QH
下载链接
链接失效反馈
官方服务:
资源简介:
<b>Description</b> <br> This dataset presents 2024 U.S. trade labels organized by HS Codes and Census long and short commodity descriptions. As part of this research and dataset contribution, user-friendly labels were developed using a large language model (LLM) to enhance readability and usability for economic research and other applications. This dataset covers nearly 30,000 unique codes traded with the U.S. in 2024. By aligning detailed trade classifications with simplified descriptions, this dataset supports more efficient trade analysis for both technical and non-technical users.<br> <hr> <b>Data Dictionary</b><br> <ul> <li>Index: Row ID, sorted by HS Code</li> <li>HS Code: HS Code (Harmonized System); HS2/4/6/10; Note: "-" indicates Total.</li> <li>Census Full Description: Commodity Description (Full - Census)</li> <li>Census Short Description: Commodity Description (Short - Census)</li> <li>Year: Trade Year</li> <li>HS Level: HS Classification Level; ; HS2/4/6/10; Note: 1 indicates Total.</li> <li>Export (0/1): Export Indicator (1 = Yes, 0 = No)</li> <li>Import (0/1): Import Indicator (1 = Yes, 0 = No)</li> <li>Country: Partner Country</li> <li>LLM Name: LLM-Generated Commodity Label</li> </ul>
提供机构:
Harvard Dataverse
创建时间:
2025-03-22
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作