Analysis of named entity recognition and linking for tweets

作者：

Highlights：

• We analyse the named entity recognition and disambiguation performance on tweets.

• Multiple state-of-the-art systems are included.

• Commercial and academic systems suffer the same range of problems.

• Lack of context is a major problem, demanding new, custom NER & NEL approaches.

• A named entity linking corpus is released with the paper.

摘要

•We analyse the named entity recognition and disambiguation performance on tweets.•Multiple state-of-the-art systems are included.•Commercial and academic systems suffer the same range of problems.•Lack of context is a major problem, demanding new, custom NER & NEL approaches.•A named entity linking corpus is released with the paper.

论文关键词：Information extraction,Named entity recognition,Entity disambiguation,Microblogs,Twitter

论文评审过程：Received 12 November 2013, Revised 16 October 2014, Accepted 22 October 2014, Available online 19 November 2014.

论文官网地址：https://doi.org/10.1016/j.ipm.2014.10.006