Document sets and folder hierarchy

The layout of a document set is extremely important when testing and configuring classification or separation. As a result, you need to create a document set and ensure that the "Include subdirectories" and "Create subfolder for each subdirectory" options are selected.

Before you create the document set however, you need to look at the actual document structure and make sure that it matches your class hierarchy. You can do this by creating a folder for each of your classes, and copying or moving your documents into their corresponding class or child class.

Important If you create a document set with a folder structure, all document subsets within that document set also have that same folder structure.

This document set hierarchy is not needed when testing extraction or using Test Validation or Test Verification, unless you want to test foldering as well.

In order to create a set of golden files for classification or separation, you need to mimic the physical layout of your project hierarchy, but you can also have multiple document sets for benchmarks. These additional sets are compared against the golden files when changes are made to your configuration settings to help you determine if the change had a positive or negative effect on your classification or separation results.