Natural Language Processing (NLP) is an interdisciplinary domain that bridges the gap between human language and computer understanding. One of the foundational tasks of NLP is to reduce the morphological variations in the text to enhance text analysis, for which stemming and lemmatization are employed. The Natural Language Toolkit (NLTK) is one of the most powerful tools in this arena. In this article, we'll delve deep into stemming and lemmatization, elucidating their differences, applications, and how to use them with NLTK.

What is Stemming?

Stemming is a heuristic process that trims the end of words with the aim of achieving the root form of a word. This process may sometimes produce non-existing words.


  • Running → Run
  • Jumper → Jump
  • Easily → Easili

What is Lemmatization?

Lemmatization is a more methodical process than stemming. It involves reducing the word to its base or dictionary form. Unlike stemming, lemmatization ensures that the resultant word is meaningful and exists in the dictionary.


  • Running → Run
  • Better → Good
  • Geese → Goose

Differences Between Stemming and Lemmatization

Basis Stemming Lemmatization
Process Trimming the word endings. Reducing words to their base or dictionary form.
Resultant Word Might not be meaningful. Always meaningful.
Accuracy Less accurate than lemmatization. More accurate as it uses vocabulary and morphological analysis.
Speed Generally faster. Might be slower due to in-depth analysis.

Using NLTK for Stemming and Lemmatization

a. Stemming with NLTK:

The NLTK library provides several stemmers, such as the Porter Stemmer and Snowball Stemmer. Here's a simple example using the Porter Stemmer:

from nltk.stem import PorterStemmer
stemmer = PorterStemmer()

print(stemmer.stem("running"))   # run
print(stemmer.stem("jumper"))    # jump

b. Lemmatization with NLTK:

For lemmatization, NLTK offers the WordNetLemmatizer, which uses the WordNet Database to look up lemmas.

from nltk.stem import WordNetLemmatizer
lemmatizer = WordNetLemmatizer()

print(lemmatizer.lemmatize("running", pos="v"))  # run
print(lemmatizer.lemmatize("geese"))             # goose

Note that in the example above, the pos argument indicates the part of speech. By default, it is set to "noun", but specifying it (like "v" for verb) can enhance accuracy.

When to Use Stemming vs. Lemmatization

  • Stemming: Preferred when speed is more essential than accuracy, for instance, in search queries.

  • Lemmatization: Ideal when the context is crucial, like in text analysis and topic modeling.


Both stemming and lemmatization serve the purpose of reducing word forms to a common base form. The choice between them should be based on the specific application, desired accuracy, and computational resources available. The NLTK library makes it simple to incorporate both techniques into NLP pipelines, aiding in improving the quality and efficiency of text analysis tasks.