Prof. Dr. Wolfgang Lehner

(TU Dresden)

"Maintaining Sample Synopsis of Evolving Datasets"

Random sampling is an appealing approach to build synopses of large data sets because random samples can be used for a broad spectrum of analytical tasks. Current research mainly considers the database static; in this setting, a sample created once remains valid for its entire lifetime. In many applications, however, such a static view is infeasible because it does not take into account the dynamic nature of the underlying data. In this talk, I will briefly summarize recent research on the problem of maintaining a random sample of an evolving dataset. As an example for the challenges of sample maintenance and the techniques required to solve them, I will discuss the problem of maintaining a random sample from a sliding window a data stream defined over a recent time interval.



Zeit: Montag, 10.12.2007, 17.15 Uhr
Ort: Gebäude 48, Raum 210