document-understanding
2020.10
false
- Getting Started
- Framework Components
- Document Classification Validation Overview
- Classification Station
- Document Classification Validation Related Activities
- Document Understanding in AI Center
- Pipelines
- ML Packages
- Data Manager
- OCR Services
- Licensing
- References
Document Classification Validation Overview
DEPRECATED
Document Understanding User Guide
Last updated Jul 29, 2024
Document Classification Validation Overview
After automatic classification, one optional (but highly recommended) step is that of classification validation.
This refers to a human review step, in which knowledge workers can review the classification results and correct them when necessary.
Using Classification Validation guarantees that, downstream, all next steps within the Document Understanding Framework and run on the correct classification information.
It is strongly recommended to use the Document Classification Validation components when:
- you are dealing with multiple document types within a single file. This is a particularly important case, because if the automatic classification is wrong, or reports too few / too many pages within a class, if data extraction is also required downstream, then the data extraction component would be applied to the wrong content.
- you need 100% accuracy in classification and there are no other ways of validating that the classification results are correct.
Validating the classification results can be done by human input through using Classification Station.
The Classification Station is available both
- as an attended activity, through using the Present Classification Station activity, or
- as Action Center tasks, through using the Create Document Classification Action and Wait for Document Classification Action and Resume activities.