This webinar is recorded and available on YouTube.
A vast amount of the information businesses generate is unstructured, usually in the form of free-text documents, images, videos, or other human-understandable formats that are not simple for machines to process. To better organize this unstructured information and to better describe content such as products and documents, many organizations have implemented hierarchical metadata structures called taxonomies. Taxonomies are a common solution in knowledge management to capture the context, meaning, and business value to help organize both unstructured information and existing data.
Taxonomies present an excellent opportunity as they are an ideal starting point for building knowledge graphs to support personalization services that improve the user experience by delivering relevant content to users. In this presentation, Connor Vilenio from Enterprise Knowledge will share his experiences with a basic approach to enhancing a taxonomy’s underlying graph model to support a recommendation system that suggests relevant content from a semantic model of the organization’s understanding of semantic relevancy.
The presentation will:
- Consider common use cases for pursuing a recommender system and how taxonomies are an ideal starting point for calculating relevancy
- Demonstrate the process for enriching taxonomies with custom relationships and polyhierarchy to convert them into basic knowledge graphs, using Ontotext GraphDB as a foundation for driving content recommendations
- Explore how an organization can derive meaningful features from the structure of their graphs to model and then calculate semantic relevance
- Describe an algorithm design that harnesses these features that is both configurable and explainable
Who is this webinar for:
- Knowledge Managers, Content Managers, and IT professionals
- Consultants, Solution Architects, and Knowledge Engineers
Expected duration:
- 45 minutes presentation
- 15 minutes Q&A session
