Login
Talks & Seminars
Title: Outlier Detection for Information Networks
Mr. Manish Gupta, University of Illinois at Urbana-Champaign
Date & Time: January 9, 2013 14:00
Venue: SIC301, 03rd Floor, C Block, Department of Computer Science & Engineering, Kanwal Rekhi Building
Abstract:
Network analysis and anomaly detection are two important subareas of data mining. The study of networks has emerged in diverse disciplines as a means of analyzing complex relationship data. Outlier (or anomaly) detection is a very broad field and has been studied in the context of a large number of application domains. For this talk, we will focus on temporal homogeneous networks and compute outliers with respect to the evolution in community membership of the network nodes. Usually objects form multiple communities, and most of the objects belonging to the same community follow similar patterns of evolution. However, there are some objects which evolve in a very different way relative to other community members, and we define such objects as temporal community outliers. One of our studies is related to finding such outliers given two snapshots of a network (Evolutionary Community Outliers (ECOutliers)) while the other study is more general and focuses on a setting of multiple network snapshots (Community Trend Outliers (CTOutliers)). In both the studies, temporal patterns are discovered and then outliers are discovered based on such patterns in an outlier-aware manner. The major challenge lies in performing cluster matching across snapshots so as to obtain temporal patterns. Another challenge is to define the outlier score once the patterns have been discovered. We propose algorithms and demonstrate the effectiveness of our algorithms in finding such outliers using both synthetic and real datasets.
Speaker Profile:
Manish received his Masters in Computer Science under the guidance of Soumen Chakrabarti from IIT Bombay, in 2007. He worked for Yahoo! Bangalore for two years. Since 2009, he has been working towards his Ph.D. with Dr. Jiawei Han at the Department of Computer Science, University of Illinois at Urbana-Champaign. He has interned over summers at Microsoft Research, IBM Research and NEC Labs America. His research interests are in the areas of data mining and information retrieval. Specifically his recent interests are in the area of mining of information networks. He has published more than 20 research papers in referred journals and conferences, including KDD, PKDD, SDM, WWW conferences.
List of Talks

Webmail

Username:
Password:
Faculty CSE IT
Forgot Password
    [+] Sitemap     Feedback