Identifying Malicious Actors on Social Media

by Srijan Kumar, Francesca Spezzano and V.S. Subrahmanian

Online social media platforms are severely compromised by the existence of malicious actors such as bots on Twitter, vandals on Wikipedia, fake accounts on Facebook, trolls on Twitter and Slashdot, and spammers who seem to be omnipresent. This tutorial presents methods to identify malicious actors in at least 4 settings: Twitter, Facebook, Slashdot, and Wikipedia. We will look at 4 broad categories of methods: (i) network based techniques where the structure of the social network is used, (ii) text based methods where the linguistic content of posts is examined, (iii) behavior-based methods which study actions of users, and (iv) real-time processes which enable defenders of social media to keep a step ahead of malicious actors. The tutorial will identify commonly used features for classifying actors into malicious vs. benign and will give a brief explanation of different algorithms both specific to social platforms and general algorithms that are platform neutral.


The slides can be downloaded here.


@inproceedings{kumar2016badactorstutorial, author = {Kumar, Srijan and Spezzano, Francesca and Subrahmanian, V.S.}, title = {Identifying Malicious Actors on Social Media}, booktitle = {Advances in Social Networks Analysis and Mining (ASONAM), 2016 IEEE/ACM International Conference on}, year = {2016}, organization = {IEEE}, }


Bias and Deserve code, WWW 2011. [Paper]

Signed network (Slashdot) troll detection code, ASONAM 2014. [Paper]

LockInfer code, PAKDD 2014. [Paper]

CatchSync code, KDD 2014. [Paper]

VEWS code, KDD 2015. [Paper]

SpEagle code, KDD 2015. [Paper]

NFS code, PKDD 2015. [Paper]

BIRDNEST code, SDM 2016. [Paper]

Please contact us if you would like your bad actor detection code linked here!


TAMU Honeypot dataset, SIGIR 2010. [Paper]

Slashdot Trolls dataset, ASONAM 2014. [Paper]

Twitter CatchSync dataset, KDD 2014. [Paper]

ASU Honeypot dataset, ICDM 2014. [Paper]

UMDWikipedia dataset, KDD 2015. [Paper]

Wikipedia Hoax dataset, WWW 2016. [Paper]

Please contact us if you would like your bad actor detection dataset linked here!