I am trying to understand if there is capability baked into RASA NLU/Core to pick up swear words or indecent words that are not good for public reading .
Is there a current way to identify this ? If yes how ? If no , what are the available options
Thank you
I haven’t checked out rasa X yet, but in the prevision version you can build a custom Component for in your nlu pipeline.
See for an example Enhancing Rasa NLU models with Custom Components – Rasa Blog – Medium
You can use a sentiment analyzer, but you could also write a function which checks for the words in the input message if they are swearwords