Articles → NLP → Chunks In NLP

Chunks In NLP

This article describes chunks in NLP.

What Is Chunking?

In NLP (Natural Language Processing), chunking is the process of grouping words into meaningful chunks—usually phrases like noun phrases (NP), verb phrases (VP), etc.

Example

Consider the following sentence.

The quick brown fox jumps over the lazy dog

The table after tokenization and POS.

Token	POS
The	DT
Quick	JJ
Brown	JJ
Fox	NN
Jumps	VBZ
Over	IN
The	DT
Lazy	JJ
Dog	NN

After chunking

[The quick brown fox] (NP)
[jumps] (VP)
[over the lazy dog] (PP)

Noun Phrase

A noun phrase is a group of words centred around a noun (the head of the phrase), along with its modifiers (adjectives, determiners, etc.). Like the tall building → Head = building (noun), Modifiers = the, tall.

Verb Phrase

A verb phrase is the predicate part — it contains the main verb and everything attached to it.

In this sentence, jumps over the lazy dog = VP

Main verb = jumps
Complement (extra info) = over the lazy dog (a PP)

Prepositional Phrase (PP)

A prepositional phrase starts with a preposition and ends with its object (a noun phrase).

Here, over the lazy dog = PP

Preposition = over
Object (NP) = the lazy dog

Prerequisite

To begin with, install the following packages.

python.exe -m pip install --upgrade numpy pandas
python.exe -m pip install --upgrade numpy pandas scikit-learn

After the installation, download the following packages.

punkt
averaged_perceptron_tagger

Example

Consider the following example: -

import nltk
from nltk import pos_tag, word_tokenize, RegexpParser

# Sample sentence
sentence = "The quick brown fox jumps over the lazy dog"

# Tokenize and POS tagging
tokens = word_tokenize(sentence)
tagged = pos_tag(tokens)

# Define a simple grammar for noun phrases (NP)
grammar = "NP: {<DT>?<JJ>*<NN>}"

# Create a chunk parser
chunk_parser = RegexpParser(grammar)

# Parse the sentence
tree = chunk_parser.parse(tagged)
print(tree)

Output

Posted By -	Karan Gupta

Posted On -	Tuesday, September 2, 2025

Query/Feedback

Your Email Id		**

Subject		*

Query/Feedback	Characters remaining 250	**