DeBot is real-time bot detection system. The project started on Feb 2015 and it has been collecting data since Aug 2015. High correlation in activities among users in social media is unusual and can be used as an indicator of bot behavior. DeBot identifies such bots in Twitter network. Our system reports and archives thousands of bot accounts every day. DeBot is an unsupervised method capable of detecting bots in a parameter-free fashion. In March 2017, DeBot has collected over 730K unique bots. Since we are detecting and archiving Twitter bots on a daily basis, we can offer two different service: bot archive API and on-demand bot detection platform.
Four published papers of DeBot:
Nikan Chavoshi, Hossein Hamooni and Abdullah Mueen, to appear In the Proceedings of World Wide Web Conference (Companion Volume), pp. xxx-xxx, WWW 2017.
Nikan Chavoshi, Hossein Hamooni and Abdullah Mueen, to appear In the Proceedings of World Wide Web Conference (Companion Volume), pp. xxx-xxx, WWW 2017.
Nikan Chavoshi, Hossein Hamooni and Abdullah Mueen, to appear In the Proceedings of the 8th International Conference on 6th IEEE International Conference on Data Mining, pp. xxx-xxx, ICDM 2016.
Nikan Chavoshi, Hossein Hamooni and Abdullah Mueen, to appear In the Proceedings of the 8th International Conference on Social Informatics, pp. 14-21, SOCINFO 2016.
An example of two correlated Twitter users. The video actual duration is 2 hours and was recorded on March 2015.