Search code examples
nlpnltkspacyrasa-nlu

NLP: Rule Based vs Machine Learning

I am trying to build an agent/chatbot.

The human will be contacting the agent using a certain phraseology (very well specified).

I am new to this field, and I discovered that I can use Rule-based or machine learning. I have read that Rasa is not a rule based framework, but at the same time it uses Spacy.

Can somebody explain this to me.

Solution

  • You can configure Rasa to use SpaCy for e.g. entity extraction, so they're not mutually exclusive. If you want to build a virtual assistant (chatbot) then Rasa will provide you far more capabilities than raw SpaCy since it is intended to be a chatbot building framework, whereas SpaCy is a general NLP library.