ENBIS: European Network for Business and Industrial Statistics

Keynote by Emmanuel Yashchin: Monitoring Massive Data Streams

24 September 2008, 10:20 – 11:10

Emmanuel Yashchin, IBM Research

In today's business environment large volumes of data are collected
and analyzed on a routine basis. Ability to process data automatically and
extract information efficiently is considered a key competitive advantage. One
of the most important problems spawned by these developments is timely detection
of emerging trends in data streams that are driven by abrupt changes in time.
Such changes could correspond, for example, to onset of shifts or drifts in
parameters of the data model. In many situations, one has no a-priori information
that would enable to predict time points or magnitudes of changes. This
talk discusses methods for monitoring this type of data, with emphasis
on the problems of detection of unfavorable changes, estimation of the
current level of parameters and retrospective data analysis.
We will also discuss a number of applications.

Return to programme