Detecting and Revealing Malicious Retweeter Groups
Class
Article
Graduation Year
2020
College
College of Engineering
Department
Computer Science Department
Faculty Mentor
Dr. Kyumin Lee
Presentation Type
Oral Presentation
Abstract
Retweeting/sharing action has enabled information to be cascaded to distant nodes on social network. Unfortunately, malicious users as a group have taken advantage of the retweeting function with coordinated behavior to falsely distort the volume of specific keywords, topics or URLs for promotional purposes (e.g., increasing public visibility of products or services). Unfortunately, little is known about their retweeting behavior as a group and how to detect them based on group-based signals. To fill the gap, in this paper, we (i) propose Attractor+ algorithm to extract retweeter groups based on the similarity of their retweeting behavior; (ii) analyze underlying characteristics of malicious retweeter groups; (iii) propose group-based features to catch synchronized and coordinated behavior; and (iv) build a predictor to classify if a group is malicious. Experimental results show that our proposed method achieved 0.954 AUC, and improved the accuracy by 19.6~42% against existing malicious retweeter detection approaches, indicating the effectiveness of our approach.
Location
Room 154
Start Date
4-13-2017 3:00 PM
End Date
4-13-2017 4:15 PM
Detecting and Revealing Malicious Retweeter Groups
Room 154
Retweeting/sharing action has enabled information to be cascaded to distant nodes on social network. Unfortunately, malicious users as a group have taken advantage of the retweeting function with coordinated behavior to falsely distort the volume of specific keywords, topics or URLs for promotional purposes (e.g., increasing public visibility of products or services). Unfortunately, little is known about their retweeting behavior as a group and how to detect them based on group-based signals. To fill the gap, in this paper, we (i) propose Attractor+ algorithm to extract retweeter groups based on the similarity of their retweeting behavior; (ii) analyze underlying characteristics of malicious retweeter groups; (iii) propose group-based features to catch synchronized and coordinated behavior; and (iv) build a predictor to classify if a group is malicious. Experimental results show that our proposed method achieved 0.954 AUC, and improved the accuracy by 19.6~42% against existing malicious retweeter detection approaches, indicating the effectiveness of our approach.