Abstract. This paper reviews the results of a study into combining evidence from multiple streams of surveillance data in order to improve timeliness and specificity of detection ...
We propose two new improvements for bagging methods on evolving data streams. Recently, two new variants of Bagging were proposed: ADWIN Bagging and Adaptive-Size Hoeffding Tree (...
Albert Bifet, Geoffrey Holmes, Bernhard Pfahringer...
Entropy has recently gained considerable significance as an important metric for network measurement. Previous research has shown its utility in clustering traffic and detecting ...
In this chapter we consider a cell phone network as a set of automatically deployed sensors that records movement and interaction patterns of the population. We discuss methods fo...
Data Stream Management Systems are useful when large volumes of data need to be processed in real time. Examples include monitoring network traffic, monitoring financial transacti...
Theodore Johnson, S. Muthukrishnan, Vladislav Shka...