水力发电2024，Vol.50Issue(10)：28-32,5.
基于ChatGPT的网络空间中水文文本信息抽取技术
Hydrological Object Information Extraction Technology in Cyberspace Based on ChatGPT
摘要
Abstract
In the cyberspace,hydrological data is characterized by large volume,variety,widespread distribution,low value density and inconsistent quality,but the objectified data organization can effectively represent the geographical spatial location,characteristic attributes and temporal attributes of hydrological objects.Applying ChatGPT for objectified information extraction from large-scale hydrological texts is an effective approach in the hydrology and water resources field to fully utilize the rich network information resources.This method can automatically extract six characteristics such as element type,element attribute,temporal scope,temporal scale,spatial scope and spatial scale from three dimensions of time,space and element.Research indicates that the rational application of ChatGPT can significantly reduce the manpower costs and lower the technical barriers in information extraction,and it preliminarily verifies the effective application of generative AI in the field of hydrology and water resources.关键词
信息抽取/ChatGPT/水文数据/文本信息/人工智能Key words
information extraction/ChatGPT/hydrological data/hydrological information/artificial intelligence分类
天文与地球科学引用本文复制引用
龚世诚,艾萍,熊传圣,宋艳红..基于ChatGPT的网络空间中水文文本信息抽取技术[J].水力发电,2024,50(10):28-32,5.基金项目
江苏省重点研发项目(BE2020729) （BE2020729）
