Outskewer: Using Skewness to Spot Outliers in Samples and Time Series

Sébastien Heymann, Matthieu Latapy, Clémence Magnien

Proceedings of the 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012), Istanbul, Turkey

Finding outliers in datasets is a classical problem of high interest for (dynamic) social network analysis. However, most methods rely on assumptions which are rarely met in practice, such as prior knowledge of some outliers or about normal behavior. We propose here Outskewer, a new approach based on the notion of skewness (a measure of the symmetry of a distribution) and its evolution when extremal values are removed one by one. Our method is easy to set up, it requires no prior knowledge on the system, and it may be used on-line. We illustrate its performance on two data sets representative of many use-cases: evolution of ego-centered views of the internet topology, and logs of queries entered into a search engine.

Download

Detecting Events in the Dynamics of Ego-centered Measurements of the Internet Topology

Assia Hamzaoui, Matthieu Latapy and Clémence Magnien

Proceedings of International Workshop on Dynamic Networks (WDN), in conjunction with WiOpt 2010

Detecting events such as major routing changes or congestions in the dynamics of the internet topology is an important but challenging task. We explore here a top-down approach based on a notion of statistically significant events. It consists in identifying statistics which exhibit a homogeneous distribution with outliers, which correspond to events. We apply this approach to ego-centerd measurements of the internet topology (views obtained from a single monitor) and show that it succeeds in detecting meaningful events. Finally, we give some hints for the interpretation of such events in terms of network events.

Download