When classification accuracy is not enough: Explaining news credibility assessment

作者：

Highlights：

• Web browser extension for online news credibility assessment.

• A visual interface for enhanced understanding of methods’ inner workings.

• Deep learning classifier compressed for deployment in a web browser.

• Users more accurate in fake news detection when supported by the tool.

• Model interpretability more important than accuracy in user studies.

摘要

•Web browser extension for online news credibility assessment.•A visual interface for enhanced understanding of methods’ inner workings.•Deep learning classifier compressed for deployment in a web browser.•Users more accurate in fake news detection when supported by the tool.•Model interpretability more important than accuracy in user studies.

论文关键词：Visual analytics,Credibility,Text classification,Fake news,Natural language processing

论文评审过程：Received 19 January 2021, Revised 30 April 2021, Accepted 25 May 2021, Available online 12 June 2021, Version of Record 12 June 2021.

论文官网地址：https://doi.org/10.1016/j.ipm.2021.102653