Investigating Entropy for Extractive Document Summarization

作者：

Highlights：

• Unsupervised algorithm for extractive single document summarization.

• Exploits entropy of semantic units in latent space of the document.

• Evaluated for three domains, two Indian languages and three European languages.

• Public data-sets used for evaluation.

• Code and data is available on github account of first author.

摘要

•Unsupervised algorithm for extractive single document summarization.•Exploits entropy of semantic units in latent space of the document.•Evaluated for three domains, two Indian languages and three European languages.•Public data-sets used for evaluation.•Code and data is available on github account of first author.

论文关键词：Entropy,Non-negative Matrix Factorization,Extractive summarization,Semantic similarity,Language independent

论文评审过程：Received 12 May 2020, Revised 21 July 2021, Accepted 26 August 2021, Available online 8 September 2021, Version of Record 24 September 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115820