Articles → NLP → NLP Pipelines

NLP Pipelines

This article describes NLP pipelines.

Purpose

In Natural Language Processing (NLP), a pipeline is a sequence of steps (modules or stages) that transform raw text into a structured representation that machines can analyse.

NLP Pipeline Stages

Text Input

Raw text data from documents, web pages, chats, etc.

Text Preprocessing

Method	Description
Tokenization	splitting text into words, sub-words, or characters.
Lowercasing	making text uniform.
Stop-word Removal	removing frequent but uninformative words (the, is, and).
Stemming / Lemmatization	reducing words to their base form.
Handling punctuation, special characters, numbers, emojis
Normalization	expanding contractions (don't → do not), correcting spelling

Feature Extraction / Representation

Converting text into numerical form:

Bag of Words (BoW)
TF-IDF
Word Embeddings (Word2Vec, GloVe, FastText)
Contextual embeddings (BERT, GPT, etc.)

Modelling / Learning

Applying machine learning or deep learning models.

Postprocessing

Converting predictions into human-readable output.

Posted By -	Karan Gupta

Posted On -	Monday, September 15, 2025

Query/Feedback

Your Email Id		**

Subject		*

Query/Feedback	Characters remaining 250	**