How to use multiple intent classifier in Rasa nlu pipeline

Rasa Open Source
1

Hello everyone: I want to use multiple intent classfier in rasa nlu pipeline,but not work,The following is my configuration file:

language: "en"

pipeline:
- name: "SpacyNLP"
- name: "SpacyTokenizer"
- name: "RegexFeaturizer"
- name: "SpacyFeaturizer"
- name: "CRFEntityExtractor"
- name: "EntitySynonymMapper"
- name: "KeywordIntentClassifier"
- name: "SklearnIntentClassifier"
- name: "DucklingHTTPExtractor"
  url: "http://localhost:8000"
  dimensions: ["time", "duration", "phone-number", "number", "amount-of-money", "distance"]

I added two intent classifiers ，KeywordIntentClassifier and SklearnIntentClassifier, but KeywordIntentClassifier not work,where is the problem？

I want to use both keyword and machine learning to identify the intent at the same time. First, the intent is identified by the keyword. If the keyword recognizes the intent, the result is returned. Otherwise, the intent is identified through machine learning.

How can I do it? Thanks a lot!

2

I don’t think this order of components will work, as the sklearn intent classifier will always overwrite the keyword result with its own prediction.

Can you try putting the KeywordIntentClassifier after the SklearnIntentClassifier in your pipeline?

3

Thanks a lot ,it works

4

perhaps you could use 2 different models for this and merge the results?

5

I’m new to RASA, can you please tell me what’s the use of having multiple intent classifiers?