Routing specific document types to avoid Knowledge Processing

Some document types are not suitable for Knowledge Processing. For example, handwritten correspondence should not be classified by Knowledge Processing. Trying to "learn" classification of such document in Knowledge Processing creates many active layout data point (LDPs), which can negatively affect overall classification performance.

Use this procedure to set up such documents to circumvent Knowledge Processing:

1. Recommended: Set up classifiers for that document type using the Capture Components Administration module (see Capture Components Administration Help for instructions). Then reimport the solution.

2. Using the Process Designer, use Split Documents and Merge Documents (see XBOUND Help) to circumvent Knowledge Processing. Example:

   

3. Set up a process rule that looks something like this:

   

   1. As the Operand, select Field value. The Field for Rule Expression dialog is displayed.

      • As the Document type name, select the XBOUND document type that should bypass Knowledge Processing.

      • As the Field name, select the classification field name that you specified when you imported the solution (see step 6 of Importing solutions from Capture Components Administration).

   2. Set Operator to equal.

   3. Set Value to the name of the document specification from the Administration module.

   4. Click Extended Rule Settings and select Rule applies to root and sub-documents and (Route if) at least one document meets the criteria:

      

4. Close the dialog and save the process.

Related topics

Creating an XBOUND process that uses Knowledge Processing

Knowledge Processing: Overview