PROJECT TITLE :

TargetVue: Visual Analysis of Anomalous User Behaviors in Online Communication Systems

ABSTRACT:

Users with anomalous behaviors in on-line communication systems (e.g. email and social medial platforms) are potential threats to society. Automated Anomaly Detection based mostly on advanced Machine Learning techniques has been developed to combat this issue; challenges remain, though, because of the difficulty of obtaining proper ground truth for model training and analysis. So, substantial human judgment on the automated analysis results is typically required to better change the performance of anomaly detection. Unfortunately, techniques that enable users to understand the analysis results additional efficiently, to create a assured judgment regarding anomalies, and to explore information in their context, are still lacking. During this paper, we tend to propose a completely unique visual analysis system, TargetVue, which detects anomalous users via an Unsupervised Learning model and visualizes the behaviors of suspicious users in behavior-rich context through novel visualization designs and multiple coordinated contextual views. Particularly, TargetVue incorporates 3 new ego-centric glyphs to visually summarize a user's behaviors that effectively present the user's communication activities, features, and social interactions. An efficient layout technique is proposed to put these glyphs on a triangle grid, which captures similarities among users and facilitates comparisons of behaviors of different users. We demonstrate the ability of TargetVue through its application in an exceedingly social bot detection challenge using Twitter information, a case study primarily based on email records, and an interview with professional users. Our evaluation shows that TargetVue is beneficial to the detection of users with anomalous communication behaviors.