Introduction to Natural Language Processing (NLP)

Jun 5, 2020

Do you know what is the common thing between Google Search, Facebook, Alexa and the US President?

It’s “Natural Language Processing”

Natural Language Processing, or NLP in short, is a technique/field of study in computer science, where we study how to teach computers to understand the text.

What is a Natural Language?

A natural language is a language that has evolved naturally in humans through use and repetition without conscious planning. According to WorldAtlas,  English is the most popular natural language in the world with 1.39 billion speakers, followed by Mandarin Chinese with 1.15 billion speakers.

Most popular natural language- What is Natural Language Processing

Most popular natural language by number of speakers


As computers can only understand numbers, natural languages and texts need to be processed before we can use it for data mining or machine learning tasks. This process is called Natural Language Processing, or NLP in short. NLP allows us to teach the machines how to understand texts, or formally defined as:

Natural Language Processing is a field in Artificial Intelligence that enables machines with the ability to read and derive meaning from human languages.

Here are the 5-step process that involved in a standard NLP.


It is difficult for machines to understand the semantics and context of a document, a paragraph, or even a sentence as a whole. First, we utilize the tokenization process to break them to down their smallest semantic units call token.

For example, given a sentence “Dr. Lau is a data scientist“, we can break them down into six tokens using space as the separator. We also remove punctuations and symbols. This approach is suitable for any language that uses blank space as a separator. For languages that do not, such as Chinese, Japanese, and Thai, we will need to use a dictionary or apply external knowledge to split words into tokens.

NLP Tokenization - What is Natural Language Processing

Tokenization breaks an English sentence into tokens using space


Stop Word Removal

The next step after tokenization is to remove stopwords. Stopwords refer to terms that appear frequently in a natural language (e.g. is, am, a, in) but do not aid in understanding the meaning. Removing stopwords prevents high occurrence words from taking up space in our database and taking up the valuable processing time.

What is Natural Language Processing

Common stopwords in the English language.


In English, a verb can appear in many different forms. For example, in an article about Internet connection, the word “connect” may appear as: connects, connected, connecting, connection, and connections. Without any processing, machines will consider these words as five different words and calculate separately. Therefore, we will need to reduce these variants into their root form “connect“. This process is called stemming.  Stemming reduce redundancy for words that have the same meaning, so that we can obtain an accurate calculation and increase the accuracy of subsequent tasks. Different stemming algorithm has their own rules, and a common stemmer for English is Porter Stemmer.

What is Natural Language Processing

Stemming reduces word variations back to its root form


In some cases, stemmer is unable to reduce word variations to its root form by simply chopping away the suffixes. Especially irregular verb pairs such as “begin, began“, “drank, drunk“, “eat, ate“, or words like “rats, mouse, mice“. We will need to utilize a dictionary or external knowledge to group inflected forms of a word together, and this process is called lemmatization.

What is Natural Language Processing

Using lemmatization to group inflected forms of a word


Part-of-speech tagging

Lastly, we perform part-of-speech tagging, also known as POS tagging, to identify the role of a word in a sentence. This is based on the relationship and context with the surrounding words. Part of speech includes nouns, verbs, adverbs, adjectives, pronouns, prepositions, and other sub-categories.

part-of-speech-What is Natural Language Processing

Part of speech (or POS) tagging marks the role of a word in a sentence


Natural Language Processing is the fundamental tasks in all text processing tasks, in order to transform unstructured text into a form that computers understand. From this point on, we can use it to generate features and perform other tasks like named entity extraction, sentiment analysis, topic detection. With the help of deep learning algorithms, we can expect NLP to reach a level of advancement in the coming years, that make complex applications such as automatic translation, human-computer interaction, more possible.

Enjoyed reading this? do share with us your thoughts in the comment section below.

If you are interested to learn more about machine learning, AI, Big data, or if you aspire to become a data scientist, LEAD offers a complete Data Science course in Malaysia that is designed to equip you for a career in data science within 8 weeks.

What are some other NLP techniques do you use? Leave them in the comment section below.

Don’t forget to visit us on Facebook and subscribe to our YouTube channel for more interesting content too.

Brought to you by

The LEAD team believes in growth. The question we ask every day: How can we help our students achieve more?

Get the scoop on the latest stuff.


Ultimate Guide to Data Science

Recent Posts

1 Comment

  1. Harsh Patel

    Hello, Amazing information shared and it’s very impressive and so much helpful for me. Keep it up and Thank you very much.:)



  1. Build your data science career in malaysia by 2021 - LEAD - […] all these unstructured data can be only analysed by using NLP. You need to know how to crawl data,…

Submit a Comment

Your email address will not be published. Required fields are marked *