liveBook · Manning

CountVectorsFeaturizer

As explained in Natural Language Processing in Action, Second Edition

CountVectorsFeaturizer is a Rasa pipeline component that converts text into count-based token features.

CountVectorsFeaturizer converts chatbot utterances into bag-of-words vectors for downstream intent recognition. It belongs to Rasa’s NLU pipeline, alongside tokenizers and other feature generators. section 12.5.1

Word and character features

A Rasa pipeline can configure one CountVectorsFeaturizer for ordinary token features and another with analyzer: char_wb, min_ngram: 1, and max_ngram: 4 for character n-grams. DIETClassifier then consumes the resulting features.

How Natural Language Processing in Action, Second Edition covers it

How deep it goes
Defines its bag-of-words role and shows configuration variants, but does not explain vector construction internals.
Examples
Uses separate word and character n-gram configurations in a Rasa chatbot pipeline.
What it assumes
Assumes familiarity with Rasa pipeline configuration and introduces count vectorization as a bag-of-words feature step.
What you can do after
Configure a Rasa NLU pipeline with word-based or character-based count features for intent classification.
How it connects
Connects tokenization, regular-expression features, lexical-syntactic features, DIETClassifier, and pretrained embeddings.
Cover of Natural Language Processing in Action, Second Edition
liveBookmanning.com
Read Natural Language Processing in Action, Second Edition on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Natural Language Processing in Action, Second Edition. The section links point to the parts of the book it draws on.