Taxonomy Design Best Practice for Knowledge Graphs
A Talk by Heather Hedden (Independent consultant, Hedden Information Management)
About this Talk
Ontologies form the semantic framework for linking data within knowledge graphs, but users often start their queries with subjects, which they may describe inconsistently. This is where a taxonomy is useful: bringing together synonyms and other variant names and arranging concepts in user-friendly browsable hierarchies or facets.
A taxonomy, whether considered part of an ontology or connected to an ontology, is thus an important part of a knowledge graph. Furthermore, taxonomy concepts are designed and implemented to be tagged to content, thus extending the scope of a knowledge graph to include not just data but also varied relevant content (documents, media, etc.)
While taxonomies are easier to design and create than ontologies, too often they are created without any skill or training. In other cases, taxonomies originally designed for a different purpose are inappropriately reused. Poorly designed or inappropriate taxonomies yield poor results.
This tutorial will cover the basics and best practices in taxonomy design, including standards, sources for topical concepts, wording of labels, alternative labels, hierarchical and associative relationships, and governance. How taxonomists connect to ontologies will also be discussed.
Goals
- Recognize where and when taxonomies are needed.
- Know what resources to use in developing or editing a taxonomy.
- Know the basics of creating good taxonomies or modifying existing taxonomies to enhance their knowledge graphs.
The problem is that including well-designed taxonomies within knowledge graphs is often overlooked.
This is important because, without a good taxonomy in a front-end application, it is difficult for users to explore data in a knowledge graph and linked content from a topical aspect, and related content may not get tagged and included correctly.
Key Topics
- Introduction to taxonomies and other types of controlled vocabularies
- Standards and models for taxonomies
- Sources for taxonomy concepts
- Wording of concept labels and alternative labels
- Taxonomy hierarchical and associative relationships
- Taxonomy and ontology comparisons and connections
Target Audience
- Ontologists or knowledge engineers who are not experienced in creating taxonomies
- Those who have a basic understanding of taxonomies or ontologies, but would like to know more
- Managers of data, information, content, or knowledge
Goals
- Recognize where and when taxonomies are needed.
- Know what resources to use in developing or editing a taxonomy.
- Know the basics of creating good taxonomies or modifying existing taxonomies to enhance their knowledge graphs.
Session outline
- Introduction to taxonomies and other types of controlled vocabularies
- Standards and models for taxonomies
- Sources for taxonomy concepts
- Wording of concept labels and alternative labels
- Taxonomy hierarchical and associative relationships
- AI and LLMs in taxonomy development
- Taxonomy and ontology comparisons and connections
- Tools for managing combined taxonomies-ontologies
Format
Most of the class will be lecture, along with ample Q&A and discussion.
There will be a few brief interactive exercises (to be entered into the online discussion chat) for participants to create alternative labels for concepts and create hierarchical relationships.
Level
Beginner - Intermediate
Prerequisite Knowledge
Basic familiarity and understanding of ontologies and taxonomies, but prior experience creating them is not required.
About The Speaker
Independent consultant, Hedden Information Management
Heather Hedden is an independent consultant specializing in taxonomies. She previously worked as a Senior Consultant at Enterprise Knowledge, as a Knowledge Engineer at Semantic Web Company, and as a Senior Vocabulary Editor at Gale/Cengage, and in other taxonomist roles.
We’re a cloud tech company that provides organisations around the world with computing infrastructure and software to help them innovate, unlock efficiencies and become more effective. We also created the world’s first – and only – autonomous database to help organise and secure our customers’ data.