An administrator supporting a highly distributed system of different workstations could be easily confronted with maintenance of a high number of separate knowledge bases, created at different locations, at different times. Thus in some cases it might be required to merge the learn sets (created in context of the same production global SLW project) from different project files, being able to provide some additional considerations.
Use this tool to quickly merge multiple SLW sub-projects, overtaking the “SLW knowledge” without using the learn set manager tool. The Merge SLW Learn Sets tool enables you for example to quickly correct corresponding configuration errors related to “disconnected” local Advanced Verifier workstations.
The tool works only with Supervised Learning Workflow (SLW) projects, which are the instances of the same master (global) SLW project. In this connection, only the SLW vendor classes can be populated and/or updated through this tool. The tool cannot update classes if the project contains any classes that are not created automatically by SLW.
Example
In the example above, the tool analyzed and found out the following.
- Base class “Invoices” with no documents in learn sets (neither source, nor target)
- Lines 2,3,4,5 the derived classes with names and number of already learned documents (both in source and target projects)
- As of line 4 the “Worldnet 1234565” has one document in the source project, but no documents in the target project, therefore the tool automatically suggests to make a copy from source to target project. You can also manually select some other nodes; for example, you’d choose two extra nodes “Unicom” and “BOLZ”, the class “Unicom” of the target project is trained with one more document available for this class in the source project and the class “BOLZ” – with two more documents available in the source project.