Recommended approach to Intent Training


(Madhu) #1

Tips when training bot for Intent detection.

  • If there are good number of sample utterances, try training the bot using machine learning approach first before trying to train the fundamental meaning model.

  • Define bot synonyms to build a domain dictionary. (example, ‘pwd’ for password; ‘SB’ for savings bank account)

  • After every change to the model training, run the batch testing modules. Test-suites are means to perform regression testing of your bot’s ML model.

    • An approach to bot NLP training is to first create a test suite of most of the use cases(user utterances) that the bot needs to identify, run it against the model and start training for the ones that failed.
    • Create/update batch testing modules for high usage utterances.
  • Publish the trained model only after detailed testing.

  • When naming the intent ensure that the name is relatively short (3-5 words) and does not have special characters, words from the stop word list. Try to ensure the intent name is close to what the user will request in his utterance.