A Natural Language Processing System for the Efficient Updating of Highly Curated Pathophysiology Mechanism Knowledge Graphs
This repository provides the code for our research regarding the extension of the Tau subgraph within the pharmacome using NLP (doi: https://doi.org/10.1016/j.ailsci.2023.100078 )
The different dataset and KG used in this study are located in data directory. These files include:
-) The triples from the original Tau subgraph of Human Brain Pharmacome (HBP), which focuses on the context of pTau
-) The NLP training dataset which is generated by combining several published KGs that were built by expert manual curation on various topics such as AD or Parkinson’s Disease.
-)The high-quality triples that are extracted by the NLP pipeline and manually evaluated.
The code directory contains the following scripts:
-)The script for the extraction of PubMed ids of the extension corpus
-)The script for analyzing the NLP-based KG.
-)The script for analyzing the training dataset