Discovering filter keywords for company name disambiguation in twitter

作者：

Highlights：

• We address the company name disambiguation task defined in WePS-3 at CLEF 2010.

• An algorithm to extract filter keywords is presented.

• The algorithm allows to classify 58% of the tweets with 75% accuracy.

• A complete classification of all tweets obtains 73% accuracy.

• Our unsupervised algorithm has a 14% loss w.r.t its supervised counterpart.

摘要

•We address the company name disambiguation task defined in WePS-3 at CLEF 2010.•An algorithm to extract filter keywords is presented.•The algorithm allows to classify 58% of the tweets with 75% accuracy.•A complete classification of all tweets obtains 73% accuracy.•Our unsupervised algorithm has a 14% loss w.r.t its supervised counterpart.

论文关键词：Twitter,Online reputation management,Name disambiguation,Filtering

论文评审过程：Available online 15 March 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.03.001