DegExt — A Language-Independent Graph-Based Keyphrase Extractor

Abstract

Abstract. In this paper, we introduce DegExt, a graph-based languageindependent keyphrase extractor,which extends the keyword extraction method described in [6]. We compare DegExt with two state-of-the-art approaches to keyphrase extraction: GenEx [11] and TextRank [8]. Our experiments on a collection of benchmark summaries show that DegExt outperforms TextRank and GenEx in terms of precision and area under curve (AUC) for summaries of 15 keyphrases or more at the expense of a non-significant decrease of recall and F-measure. Moreover, DegExt surpasses both GenEx and TextRank in terms of implementation simplicity and computational complexity.

Extraction of Statements in News for a Media Response Analysis

“…The Text Summarization method DegExt [6] is very language-independent, because the only required NLP resource is a tokenizer. DegExt allows to choose the number of keywords (referred to as N) and, as a consequence, the size of the summaries.…”

“…Turney extracts important phrases by learned rules [12], while Mihalcea and Tarau build graphs using Page Rank and a similarity function between two sentences [7]. A language-independent approach for Text Summarization proposed by Litvak et al [6] is called DegExt. The approach transforms a given text into a graph representation where words become nodes.…”

“…The extraction of relevant statements for a MRA is related to several kinds of areas: the automated creation of Text Summaries [1,6,7,12], Information Extraction [3,13] and Opinion Mining [8,9,11].…”

Keyword and Keyphrase Extraction Techniques: A Literature Review

“…Marina Litvake et al (2011) proposed DegExt, a graph-based, cross-lingual keyphrase extractor [31]. DegExt uses graph representation based on the simple graph-based syntactic representation of document and enhances the traditional vector-space model by taking into account some structural document features.…”

Keyword extraction from single documents using mean word intermediate distance

“…CRF is an undirected graphical model that encodes a conditional probability distribution with a given set of features. Litvake et al [10] proposed DegExt, which is an unsupervised, graph-based, cross-lingual keyphrase extractor. DegExt uses a graph representation based on the simple graph-based syntactic representation of text and web documents, which enhances the traditional vector-space model by taking into account some structural document features.…”