A knowledge graph of healthcare entities and relations. This graph is based on the open data standards such as UMLS and real healthcare data such as EMR.
- 8M concepts, 83M entities, and 295M relations
- A framework to construct the healthcare knowledge graph
- A tool to help annotate the healthcare records semi-automatically
- A tool set to create the healthcare knowledge graph
- A tool to visually generate and run the healthcare data analysis workflow
No | Name | Description | File Name | Size (compressed/original) |
1 | Health concept knowledge graph | From open resources and data sources such as MeSH, ICD, SNOMED etc. 8M concepts, 32M relations | ConceptKG_2019-01-15.7z | 34M/439M |
2 | Web health concept knowledge graph | From the health knowledge crawled from the Internet. 86778 entities, 14463 relations | WebHealthKG_2019-01-15.7z | 98K/943K |
3 | MIMIC-III instance knowledge graph | From MIMIC-III dataset. 46471 patients, 18702003 entities, and 198111331 relations | Mimic3KG_2019-01-15.7z | 61M/838M |
4 | C1 instance knowledge graph | One patient in C1 hospital. C1 has 49431 patients | C1Data_2019-01-15.7z | 3K/29K |
5 | C2 instance knowledge graph | One patient in C2 hospital. C2 has 2.2M patients. | C2Data_2019-01-15.7z | 3K/29K |
Note: For privacy issue, the complete C1 and C2 knowledge graphs cannot be downloaded. If you want to research on C1 and C2 knowledge graphs, pleaes contact us.
If you have any question, please contact us: [email protected] or [email protected]
The knowledge graphs are for research purpose only.