The training type defines the algorithm that is used during the training of the locator and therefore how documents are extracted. Changes you make to the locator training type become effective after you retrain the project. The locator training type does not affect or restrict the used Knowledge Bases. You can select the algorithm for the creation of the Knowledge Base when you create the Knowledge Base. For the best result, use generic Knowledge Bases and restrict the training type for newly trained documents to Specific. This combination enables the use of certified generic Knowledge Bases that you can then improve with specific document training during online learning.

The different training types include:

Generic

This type is mainly based on keywords and keyword location in relation to an alternative or field. As a result, this option requires good OCR quality to extract data from documents. Generic training can also extract values from unknown documents that use similar keywords as some of the trained sample documents.

Specific

This type learns the layout of a sample document and only applies the extraction to another document with the same layout. This is typically true for invoices from the same vendor. Because of the trained layout, specific training can use a combination of restricted keywords, fixed positions and a known field format. In combination, these provide significantly better recognition and accuracy than generic extraction. However, this type of training only works for trained layouts.

Both

If you use both training types, a field is trained for the generic and specific search method. During extraction, the results for specific search are executed first, and if a result is found, the locator returns only the results for the specific search method. If no result is found, the generic search method is executed and if a result is found, the locator returns those results.