Dataset Overview
| Attribute | Description |
|---|---|
| Dataset Name | Medical knowledge mapping data sets |
| Overwrite knowledge base | ICD-10 catalogue, ICD-11 catalogue, industry metadata mapping database, industry metadata database, medical terminology mapping database, pharmaceutical and disease mapping database, medical terminology definition database |
| Entity size | 5 million+ Medical entities |
| Triad. | 20 million+bars (Entity-Relation-Entity) |
| Terminology standards | ICD-10 ICD-11 SNOMED CT Customized industry metadata |
| Data Format | JSON-LD / RDF / CSV / Neo4j Cypher |
| Source | Authorized to operate + Self-study |
| Use Cases | Knowledge mapping • Smart body knowledge base • Model training RWE research |
| Delivery Cycle | 1-2 weeks |
Dataset fields
| Category | Field | Description |
|---|---|---|
| Entity identification | Type of entity, entity name, entity ID, corporate classification | Harmonized entity codes and type attribution |
| Relationship Description | Type of relationship, target entity, orientation of relationship, relationship weight, confidence | Inter-entity syntaxal relations tri-cluster |
| Coding System | ICD-10, ICD-11, SNOMED CT, custom code | Multi-standard medical term encoding map |
| Terminology, definitions of terms, synonyms, English equivalents, abbreviations | Standardization of terminology and discrimination | |
| Bibliography | Source literature, source database, publishing agency, version number | Documentation support for knowledge entries |
| Metadata | Metadata name, metadata interpretation, mapping target standards, data domain | Standardized governance of industry metadata |
De-sensitization JSON-LD 3-D model
Other Organiser
"@content": "https://schema.org",
"@type": "MedicalEntity",
@id: "kg: entity/disease/I21."
"entity type": "Disease",
"entity name": "acute frontal myocardial infarction",
"icd10 code": "I21.0",
"icd11 code": "BA41.0",
"snomed ct": "54329005",
"relations": Laughs
Other Organiser
"relation type": "has symptom",
"taget entity": "kg:entity/symptom/chest pain",
"target name": "the chest pain,"
"confidence": 0.98,
"source": "ICD-11 Clinical Defense"
{\cHFFFFFF}{\cH00FFFF}
Other Organiser
"relation type": "has risk factor",
"taget entity": "kg:entity/risk/hypertension",
"target name": "high blood pressure,"
"confidence": 0.95,
"source": "WHO Global Health Issues"
♪ I'm sorry ♪
♪ I'm sorry ♪Other Organiser
@id: "kg: relation/drag disase/1000342",
Subject:
"entity type": "Drug,"
"entity name": "Atophate calcium."
"atc code": "C10A05",
"approval nomber":
{\cHFFFFFF}{\cH00FFFF}
"predicate": "indicated for",
Subject:
"entity type": "Disease",
"entity name": "high cholesterol sepsis,"
"icd10 code": "E78.0"
{\cHFFFFFF}{\cH00FFFF}
"confidence": 0.99,
"Source": ["Dicts Note", "China Pharmacy 2020 Edition" ],
"Videence level": "Grade A"
♪ I'm sorry ♪Other Organiser @id: kg:mapping/term/HTN 001 "source system": "Hospital HIS custom code," "source code": "D-HYP-ESS", "source term": "Natural hypertension," "target standard": "ICD-10", "target code": "I10", "Target term": "Essentic (primary) hipertension," "target standard 2": "ICD-11", "target code 2": "BA00", "Target term 2": "Essentic Hipertension", "mapping type": "exact metch," "confidence": 0.97, "revied by": "clinical termologist." ♪ I'm sorry ♪
Other Organiser
"Meta data domain": "Medical clinic,"
"fields":
Other Organiser
"field name": "chief complaint",
"definition": "the main symptoms of a patient's visit or an inadequate description,"
"data type": "TEXT",
"mapped standard": "HL7 FHIR R4 Condiment.code,"
"Mapped path": "Condition.code.text."
{\cHFFFFFF}{\cH00FFFF}
Other Organiser
"field name": "diagnosis code",
"definition": "the ICD-10 code for the main diagnosis of discharge."
"data type": "Varchar (10)",
"mapped standard": "ICD-10,"
"mapped path": "Condition.code.coding ICD-10.code"
♪ I'm sorry ♪
♪ I'm sorry ♪Other Organiser @id: "kg:term/definition/TCM 0127", "term": "Demonstration rule," "english name": "Syndrome Difference and Treadment", "definition": "one of the basic characteristics of Chinese medicine, identifying the evidence by analysing the data of the four medical consultations and identifying the corresponding treatment principles and methods," "Domain": "The foundation of Chinese medicine," "synyonyms": ["Dividing treatment", "Demonstration",] "related terms": "Source": "The Basics of Chinese Medicine," "version": "2024.Q2", "review status": "verified." ♪ I'm sorry ♪
Knowledge labeling and quality control specifications
Entity standardization
Harmonize entity naming norms, denigration of synonyms, multi-standard code of association
Relationship verification
Medical professionals examine the clinical reasonableness of each relationship
Confidence rating
≥0.95 Grade A / 0.85-0.94 Grade B / <0.85
Version Control
Note number and audit status for each knowledge entry
AI Application Scenarios
Knowledge Injection for Medical Large Models
Pre-training and fine-tuning of LLM ' s structured medical knowledge base
Clinical decision support system
Building the CDSS Logic Engine based on the disease-symptomological-drug knowledge map
Standardized conversion of terms
HIS/EMR Custom Encoding Automaps to ICD/SNOMED Standard Encoding
Smart body knowledge base build
Provide a searchable structured medical knowledge base for medical AI Agent
Drug repositioning study
New adaptations based on drug-disease-target mapping
Improperization of medical education
The Glossary and Network of Relationships supports the adaptive medical teaching system
Commercial Value
Shorten the knowledge engineering cycle
The construction of a medical knowledge map from zero usually takes 12 to 18 months, and using this data set can reduce the cycle to 1-3 months.
Multi-standard interoperability
The three criteria ICD-10/11 and SNOMED CT are also covered, addressing the problem of semantic interoperability between hospital information systems.
Commercially Usable in Compliance
All knowledge entries are sensitized and authorized to support direct integration of commercial AI products.
Maintenance updated on an ongoing basis
Updates are synchronized with ICD/WHO code version, and at least two large versions are published annually.
Frequently asked questions
What are the core values of the medical knowledge mapping data set?
The core value is to provide a tri-partite group of 5 million+ entities with 20 million+ relationships that have been examined by clinical experts, covering full ICD-10/11 coding and SNOMED CT mapping, which can be directly injected into the large model training process and significantly lower the cost of building the knowledge engineering of medical AI.
How can the accuracy of the intellectual discourse be ensured?
All knowledge entries are reviewed by the Group of Clinical Medicine, and source literature is traceable.
Does the custom knowledge map Schema support?
Support. Data sets provide standard RDF/JSON-LD formats for export, compatible with mainstream maps such as the Neo4j/JanusGraph/ArangoDB, and customised remaps based on your operations Schema.
Complete catalogue of access to medical knowledge
Contact data experts get detailed Schema and business offers.