Skip to main content
To achieve good data extraction accuracy, we’ll use a separate set of rules for each class of documents. By default, each activity is trained using all documents uploaded to the skill. However, in our case it is more efficient to train each Extraction Rules activity using documents of a single class. So we’ll have to separate the document sets for the two activities.
  1. Enable the activities document sets.
    a. Open the “Sick Note DE” activity in the Activity Editor.
    b. Click All Documents.
    c. Select Sick Note DE Document Set from the drop-down list.
    d. Repeat steps a-c for the “Sick Note BE-NL” activity.
  2. Click the name of the skill and then go to the Documents tab. The two Document sets named after the Extraction Rules activities will appear in the list on the left.
  3. Select all German documents from the All Documents set and click Add to Set.
  4. Select “Sick Note DE” from the drop-down list.
  5. Repeat steps 3 and 4 to populate the “Sick Note BE-NL” document set.