Predictive State Representations (PSRs) have shown a great deal of promise as an alternative to Markov models. However, learning a PSR from a single stream of data generated from ...
We address the problem of computing approximate answers to continuous sliding-window joins over data streams when the available memory may be insufficient to keep the entire join...
We consider the problem of maintaining aggregates and statistics over data streams, with respect to the last N data elements seen so far. We refer to this model as the sliding wind...
Mayur Datar, Aristides Gionis, Piotr Indyk, Rajeev...
While there has been a lot of work on finding frequent itemsets in transaction data streams, none of these solve the problem of finding similar pairs according to standard similar...
Mining evolving data streams for concept drifts has gained importance in applications like customer behavior analysis, network intrusion detection, credit card fraud detection. Se...