Graphite Knowledge Studio
Enterprise Taxonomy Management
Ontology Schema Management
RESTful API Integration
Graphite simplifies the management of enterprise taxonomies and ontologies. Graphite Knowledge Studio enables taxonomists and content managers to automate the tagging and categorization of enterprise content.
Graphite Knowledge Studio gives taxonomists and content managers full control of how enterprise content is tagged and categorized through simple and transparent rules that yield explainable results.
Graphite Knowledge Studio, powered by Ontotext Text Analytics, delivers multiple benefits:
- Plug-and-play utilization of existing single source of truth enterprise taxonomies and ontologies
- Out-of-the-box tagging based on standard taxonomy semantic schema such as SKOS
- Rapid refinement of tagging accuracy by adding tagging contexts to the taxonomy
- Improve both recall and precision by harmonizing the semantics of search and content terminology
- Identify what is most relevant or salient – the aboutness of a document
- Support content recommendation and discovery
Semantic tagging and auto-categorization improve the accuracy and consistency of metadata.
Semantic metadata powers precision search and supports content recommendations and discovery.
How it works
identifies the many taxonomy concepts and named entities that are mentioned with in the full text of a document. It uses concept labels, disambiguators and contextual rules.
identifies the few concepts and named entities that best describe the aboutness of a whole document. It uses term frequency, document zone relevancy, semantic proximity, interencing, and TF-IDF.
identifies new named entities (people, places, organizations, etc and/or conceptual entities, that are found in the full text of a document and not present in the taxonomy. It uses Natural Language Processing (NLP).
Big Knowledge Graphs
such as DBpedia, GeoNames and Wikidata, can be used for disambiguation and to identify content recommendations. They use inferencing and similarity matching to identify connections between concepts and content.
Human in the Loop
regardless of whether rules-based or ML methods are employed, the process of developing and refining tagging and categorization pipelines iterative and incorporates human feedback.