Entities aren't being extracted as intented

Rasa Open Source
1

Hi,

I was following the docs on RASA. I was trying to extract entities from a particular input statement. In my case, I was trying to extract the name of the user, but it doesn’t seem to work properly, since it works only for the names which I have provided in the intent.

DOC link: https://rasa.com/docs/rasa/nlu/entity-extraction/

Intent:

 ## intent:tell_name
- My name is [john](name)
- I am [john](name)
- My first name is [John](name)
- You can call me [john](name)

Story: ## greet and ask name * greet - utter_greet * tell_name{“name”: “john”} - utter_acknowledge_name * thanks - utter_goodbye

Domain:
intents:
  - tell_name
  - thanks
entities:
  - name

slots:
  name:
    type: text

responses:

utter_greet:
  - text: "Hey! How are you?"

  utter_acknowledge_name:
  - text: "Hmm, I know your name now {name}"
  - text: "I'll call you {name} from now on :)"

So, when I type in “My name is john”, it saves it in the slot, but that doesn’t happen with a different name :frowning:

Thanks

2

some of the entity extractors such as SpacyEntityExtractor comes with build in PERSON entity that can you use to extract person names.

3

Hi, I am aware of extracting names using spacy, but it becomes inefficient when come custom entities specific to the chat bot must be extracted. Though, we have an option of using a lookup table, I am curious to know if there are methods with which custom entities can be extracted by analyzing the pattern from the input?

4

For domain specific entities you need to specify them. Lookup Tables will help you in regard to this