3 Create your first knowledge graph from ontologies
This chapter covers
- Building your first knowledge graph starting with business goals
- Selecting the best technology for the knowledge graph representation based on use cases
- Constructing a valuable knowledge graph to support clinicians' activity
The KG construction is complex because it requires extracting and integrating information from heterogeneous data sources. We can relate the difference between these sources to the data format (XML, CSV, or JSON), the storage technology (relational or document-oriented databases), and the information syntax (2022-08-09 or 9 August 2022). Another key difference involves the meaning of the data, which becomes critical for information exchange. In the case of healthcare, for instance, the adoption of diverse expressions to identify the same concept (diabetes type 2 or ketosis-resistant diabetes), the use of the same acronym to define distinct concepts (PE to identify physical examination or pulmonary embolism), and various information granularity (necrosis or lobular necrosis) represent a huge obstacle to data integration.