Am kind of confused about the term NLP…is the rasa it self NLP model to be used for creating a chatbot …or rasa uses a so called such and such NLP model inorder to train the nlu…if so what kind of NLP model does rasa uses…??
Rasa is a framework which uses state of the machine learning and NLP techniques to build the chatbot.
@ChikkaUdayaSai What kind of NLP model do rasa uses to build it??
They have several model pipelines. Out of which DIET model is very famous. Read more about that here.
@ChikkaUdayaSai…if diet classifier is a model then are the tokenzers, feuturizers are also NLP model or they are machine leaning algorithms for training out data?
They are not models. They are pre processing techniques to convert raw text into a form through which training of DIET classifier starts.
@ChikkaUdayaSai…okay wait …assume a user inputs a query saing that what can you suggest if a person has malaria? and the bot replays doing such and such will decrease your pain… I used supervised embedding such as whitespacetokenizer,regexentityextracter,regexfeuterizer,countvectorfeuterizer,entitysynonmmapper,LexicalSyntacticFeaturizer,dietclassifier…and the policy Ted policy,rulepolicy and memoization policy and how these pipeline and policy will manage the user input and the bot output?
White space tokenizer will split the sentences into individual words. regex entity featurizer will check for patterns in the data like mail id, phone number and so on. Count vector featurizer will count the occurrence of each individual word.
All these are nothing but features (input vector).
These features will become input to the Diet classifier.
Read more about these here: