Dataset for natural language processing-based historical site assessment in nuclear decommissioning

This repository comprises two main components: an aggregated dataset of nuclear power plant (NPP) reports and code developed to identify contamination-relevant information in the NPP reports. The dataset is constructed from 100 accident and failure investigation reports obtained from the Korea Institute of Nuclear Safety and 15 reports from the U.S. Nuclear Regulatory Commission (US NRC). The US NRC reports have been translated into Korean to ensure consistency. The code extracts contamination-relevant information, including associated locations and components, from historical NPP documents. Detailed descriptions of the dataset structure and the code are provided in the README text file.