Abstract. Mining of data streams must balance three evaluation dimensions: accuracy, time and memory. Excellent accuracy on data streams has been obtained with Naive Bayes Hoeffdi...
Albert Bifet, Geoffrey Holmes, Bernhard Pfahringer...
Pattern mining algorithms are often much easier applied than quantitatively assessed. In this paper we address the pattern evaluation problem by looking at both the capability of ...
Stream computing research is moving from terascale to petascale levels. It aims to rapidly analyze data as it streams in from many sources and make decisions with high speed and a...
Ankur Narang, Vikas Agarwal, Monu Kedia, Vijay K. ...
While scalable data mining methods are expected to cope with massive Web data, coping with evolving trends in noisy data in a continuous fashion, and without any unnecessary stopp...
Abstract. The data stream model of computation is often used for analyzing huge volumes of continuously arriving data. In this paper, we present a novel algorithm called DUCstream ...