Train Extractors Scope

Properties panel

• DisplayName - The display name of the activity.

• DocumentObjectModel - The Document Object Model you want to use to validate the document against. This model is stored in a Document variable and can be retrieved from the Digitize Document activity. Visit Digitize Document to learn how to achieve this. This field supports only Document variables.
• DocumentPath - The path to the document you want to validate. This field supports only strings and String variables.
  Note: The supported file types for this property field are: .png, .gif, .jpe, .jpg, .jpeg, .tiff, .tif, .bmp, and .pdf.
• DocumentText - The text of the document itself, stored in a String variable. This value can be retrieved from the Digitize Document activity. Visit Digitize Document to learn how to achieve this. This field supports only strings and String variables.
• HumanValidatedData - The output from the Validation Station, stored in an ExtractionResult proprietary variable.
  Note: If the DocumentObjectModel and HumanValidatedData files are from different sources, an error is thrown at runtime.

• Private - If selected, the values of variables and arguments are no longer logged at Verbose level.

1. Add a Train Extractors Scope activity to your workflow.
2. In the Properties panel, add your custom configurations in the DocumentObjectModel, DocumentPath, DocumentText and HumanValidatedData fields.
3. Add an extractor trainer activity (for instance, Machine Learning Extractor Trainer), or create a custom activity for extractor training by implementing the classes and place it inside the Train Extractors Scope activity.
   1. To configure your extractors, select Configure Extractors.

      You can now see the Configure Extractors wizard.

      Figure 1. Overview of the Configure Extractors wizard
      

      

      
      
   2. To use a specific extractor on a whole document type or a particular field within that document, select the check boxes next to the document type or fields. By selecting a whole document type, all corresponding fields will be automatically selected too.
   Figure 2. Animated image showing the process of selecting document types or corresponding fields for extraction
   

   

   
   
4. Select Get of refresh extractor capabilities, for the extractors that support this functionality, to map your taxonomy fields with the available extractor fields, or refresh them in case the extractor fields have changed. After you configure the capabilities, select Get Capabilities, to reflect the changes in the Configure Extractors wizard.
   Figure 3. Animated image showing the process of refreshing the extractor capabilities for a Machine Learning Extractor
   

   

   
   
5. To train the extractor based on its extraction result, start by entering any alphanumeric value in the Framework Alias field. After this, go on to set the exact value in the corresponding Framework Alias field of one or more trainers. This creates a link between the extractor and the respective trainer(s).
6. Select Save once all the extractors are selected.