Intent Hierarchy¶
The intent training module includes a Hierarchy feature that enables intent grouping and layering. This means that parent intents are created which can inherit the example sentences from child intents to provide more flexibility to the agent's understanding and increase the chances of the correct intent being identified.
Overview¶
Intents are added as children by dragging and dropping on the parent intent via the intent list.
Intents can be structured into three layers:
- Level 1 Intents
- Level 2 Intents
- Level 3 Intents
Tip: Use Intent Naming Conventions
It is recommended to follow a pattern when naming your intents. Either create a topic category name or use a numbering system to group intent families.
Level 1 Intents can inherit Example Sentences from their level 2 child Intents, which can inherit Example Sentences from their level 3 grandchild Intents.
Inheritance¶
Enable the "Inherit Example Sentences from Child Intents" slider for each intent to incorporate the example sentences of all child intents in the training model.
When an inherited example sentence is found, both the parent and the child intent will score a high intent mapping result but the child intent will be placed in the input.intent
property. Find the parent as an additional high scoring intent in the input.nlu.intentMapperResults.scores
array.
You can find the whole found Intent hierarchy in the input.intentLevel
object and in the input.nlu.intentMapperResults.intentPath
array.