This page was last edited on December 20, 2019, at 22:08.
Comments or questions about this documentation? Contact us for support!
A training data object is a category tree plus a set of text objects (such as emails), called Training Messages or Training Interactions, with each text object assigned to one category in the tree.
There are several ways that a text object can end up assigned to a category.
Creating a new Training Data Object process can take from a few seconds to many hours, depending on the number of interactions and agents included. Because of this, Genesys recommends that no more than 200–300 agents participate in the categorization process. You can create an agent group specifically for creating Training Data Objects, consisting of agents whose categorization is trustworthy, and who understand that their categorization decisions will have a direct effect on the classification model. You can then select that group in Step 3 above.
You cannot select individual agents.
You cannot delete any training data object (or any root category) that has at least one model assigned to it.
You can