TextRank
A classical unsupervised extractive text summarization algorithm inspired by PageRank. Sentences are treated as nodes and a directed graph is constructed using lexical overlap between sentence pairs; iterative computation of node voting weights ranks sentences by importance for summary extraction.
Timeline (last 90 days)
Traditional graph-based summarization methods (such as TextRank and LexRank) treat sentences as nodes and sentence similarity as edges, selecting important sentences via centrality algorithms.
TextRank draws on the PageRank algorithm, treating sentences as nodes, constructing a directed graph based on lexical overlap between sentences, and iteratively computing node weights.
All Facts (2)
Traditional graph-based summarization methods (such as TextRank and LexRank) treat sentences as nodes and sentence similarity as edges, selecting important sentences via centrality algorithms.
50%UnverifiedTextRank draws on the PageRank algorithm, treating sentences as nodes, constructing a directed graph based on lexical overlap between sentences, and iteratively computing node weights.
50%