N-gram

See also: Machine learning terms

Introduction

In the field of machine learning and natural language processing, an N-gram is a contiguous sequence of N items from a given sample of text or speech. N-grams are widely used for various tasks in computational linguistics, such as statistical language modeling, text classification, and information retrieval. The term "N-gram" is derived from the combination of the letter "N" and the word "gram," which originates from the Greek word "gramma," meaning "letter" or "written character."

Types of N-grams

N-grams can be categorized based on the value of N:

Unigrams (1-gram)

A unigram is a sequence of a single item, such as a word or character, from a given text. Unigrams provide a basic representation of a text and are the simplest type of N-grams.

Example: Text: "Machine learning is fun." Unigrams: ["Machine", "learning", "is", "fun."]

Bigrams (2-gram)

A bigram consists of two consecutive items from a given text. Bigrams are often used to capture information about word pairs and their co-occurrence.

Example: Text: "Machine learning is fun." Bigrams: [["Machine", "learning"], ["learning", "is"], ["is", "fun."] ]

Trigrams (3-gram)

A trigram is a sequence of three consecutive items from a given text. Trigrams are used to represent the context of word triplets and their relationships.

Example: Text: "Machine learning is fun." Trigrams: [["Machine", "learning", "is"], ["learning", "is", "fun."] ]

Applications of N-grams

N-grams are utilized in various applications of machine learning and natural language processing:

Language Modeling

N-grams are used in statistical language models to predict the probability of a word or sequence of words, given the context. Language models are an essential component of many natural language processing tasks, such as speech recognition, machine translation, and text generation.

Text Classification

In text classification tasks, N-grams can be employed as features to represent the content of a text document. They help in capturing the local context and word order information, which can be useful in differentiating between various classes or topics.

Information Retrieval

N-grams are used in information retrieval systems, such as search engines, to improve the effectiveness of query matching and document ranking. They can also help in dealing with issues like spelling variations and word proximity.

Explain Like I'm 5 (ELI5)

Imagine you have a bunch of Lego blocks with words written on them. An N-gram is a way to connect these blocks to make different combinations. If you connect just one block, that's a unigram. If you connect two blocks, that's a bigram. And if you connect three blocks, that's a trigram.

These N-grams help computers understand and predict what words come next in sentences, and they can be used for things like making search engines better or helping computers understand and create text.