A holistic system for troll detection on Twitter

作者：

Highlights：

• Up-to-date analysis of the state of the art for the problem of troll detection.

• Systematic collection and grouping of features, on Twitter.

• Construction of a working holistic system, with a very high accuracy (95.5%).

• Comparison among the different features, with a machine learning approach.

摘要

•Up-to-date analysis of the state of the art for the problem of troll detection.•Systematic collection and grouping of features, on Twitter.•Construction of a working holistic system, with a very high accuracy (95.5%).•Comparison among the different features, with a machine learning approach.

论文关键词：Troll detection,Machine learning,Online social networks

论文评审过程：Received 27 March 2018, Revised 20 July 2018, Accepted 5 August 2018, Available online 15 August 2018, Version of Record 20 August 2018.

论文官网地址：https://doi.org/10.1016/j.chb.2018.08.008