zoom_in Click to enlarge

TextRank

▲ ▼

description TextRank Overview

TextRank is an unsupervised, graph-based algorithm used for extractive text summarization and keyword extraction. Introduced in 2004 by Rada Mihalcea and Paul Tarau, it operates by building a graph where sentences or words represent vertices connected by their semantic or lexical similarity. The algorithm applies a variation of Google's PageRank formula to these vertices to identify the most central and representative text. It is primarily used by developers and researchers in natural language processing.

help TextRank FAQ

What is TextRank used for?

TextRank is an unsupervised graph-based algorithm for extractive text summarization and keyword extraction. It ranks sentences or words so important material can be selected from a document.

Who introduced the TextRank algorithm?

TextRank was introduced by Rada Mihalcea and Paul Tarau in 2004. Their method adapted graph-ranking ideas to language tasks.

Does TextRank generate new sentences with a language model?

No. TextRank is primarily extractive, meaning it selects important sentences or terms from the source. It does not require a neural language model to write new wording.

How does TextRank rank sentences?

It represents sentences or words as vertices in a graph and connects related items through semantic or lexical links. The graph structure is then used to identify the most important content.

swap_horiz

Looking for TextRank alternatives? Compare top competitors ranked & scored

arrow_forward