Datasets for SciEdit publication of "Improving precision in organization name recognition"
收藏官方服务：
资源简介：
A two-stage ROR anchored approach This research builds a pipeline for matching and validating researcher affiliation strings to ROR (Research Organization Registry) identifiers. The goal is to achieve accurate institution-level normalization by combining rule-based exact matching, fuzzy string matching, and geographic validation. --- ## Data Files The pipeline relies on the following datasets: - `20251220_Gold_Standard_Dataset.csv` Gold standard annotations used for performance evaluation. - `20251220_SciEdit_ROR_corpus_Submit.json` ROR institution master data used as matching candidates. - `geo_data.json` Auxiliary geographic dictionary for country and region normalization.
提供机构：
Zenodo创建时间：
2026-01-23
