Inference Tagging
Inference tagging makes it possible to annotate documents beyond explicit concepts by deriving new information from implicit concepts. Rather than having to collect additional data to extract tags, inference tagging leverages the existing dataset to generate new knowledge and connections.
Inference tagging is built on ontologies and rules. Where ontologies define the structural model of the data, including classes and subclasses, rules establish the laws that the data must adhere to. By utilizing these tools, inference tagging can reveal hidden relationships within the data, leading to valuable insights and connections that might have been previously unnoticed.
With inference tagging you can interpret the data, draw conclusions, and make predictions. It is a critical component in many automated decision-making processes, as it helps computers understand complex patterns and relationships within data.
Identify the hidden knowledge in your documents!
Implicit tagging
Scenario-based tagging
With inference tagging, the combination of concept tagging and annotation rules allows the flexible classification of documents according to changing readings or scenarios.
Multi-domain tagging
Even very different branches of a taxonomy can be used with flexible rules to annotate documents and allow multiple classifications in different domains.
Conditional tagging
Example use cases
Risk identification
To analyze and annotate reports, procedures, and statements, etc., inference tagging lets you associate risk types and risk levels to the concepts detected in the text. In the healthcare domain, for example, the application of inference tagging to patient data can facilitate the prognostication of health risks, the identification of preventative strategies, and the elucidation of potential therapy interventions.
Pattern detection
With inference tagging, you are able to act on occurence patterns of the concepts in a document or even their structural relationship. You can produce signals which may trigger a certain classification of the document and warnings, alarms or actions.
In manufacturing, this may be signals to review a process or start maintenance derived from shift reports.
Personalized recommendations
The flexible classification of documents according to changing readings or scenarios can be realized in modeling user behaviour and user preferences in the inference rule.
In the area of personal recommendation systems – e.g. in commerce, retail or self service – this enables recommendation systems that take context, user intent and user behavior into account.
Decision support
A document that has been enriched with implicit concepts and facts using inference tagging makes it possible to link further background information that is already focused on the scope of the decision.
A patient’s medical history can thus support doctors in treatment by providing links to diagnostic suggestions based on the symptoms entered by the healthcare provider.
Useful Resources
Inference Tagging Demo
Watch the chapter about Inference Tagging in our Webinar “What’s New: PoolParty Recommender”
Tagging 101
Learn more about Concept Tagging and how you can benefit from precise automated tagging and classification
Inference Tagging Workbench
Have a look into our documentation of our Workbench to see how PoolParty supports you in drafting your inferencing rules.