Assessing the impact of software on science: A bootstrapped learning of software entities in full-text papers

作者：

Highlights：

• We propose an improved bootstrapping method to extract software entities from full-text papers.

• A positive correlation is found between the number of mentions and the number citations.

• Software is widely used in the science community along with a substantial uncitedness.

• The 80/20 rule has been found in software mentions and citations.

摘要

•We propose an improved bootstrapping method to extract software entities from full-text papers.•A positive correlation is found between the number of mentions and the number citations.•Software is widely used in the science community along with a substantial uncitedness.•The 80/20 rule has been found in software mentions and citations.

论文关键词：Entity extraction,Information extraction,Software,Software citation,Citation analysis,Bootstrapping

论文评审过程：Received 27 May 2015, Revised 30 July 2015, Accepted 30 July 2015, Available online 10 September 2015, Version of Record 10 September 2015.

论文官网地址：https://doi.org/10.1016/j.joi.2015.07.012