Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

Use of models for prediction of avalanche hazards and monitoring of financial databases.

Back
Abstract:
System identification and process monitoring are branches of automatic control, which can be useful in seemingly unrelated fields like snow science or information technology. The first part of the talk is devoted to the prediction of avalanche hazards. It is shown that dynamical ARX-like prediction models are superior to the classical nearest-neighbor models, and can be further improved by application of classical data mining algorithms. Moreover, dynamical models can be useful to automatic validation of numerous sensor data. The presented results have been obtained within cooperation between Predict AG and Swiss Federal Institute for Snow and Avalanche Research in Davos. The second part of the talk demonstrates how selected process monitoring and system identification methods can be applied for data quality monitoring in huge financial databases. The first step to such monitoring is data aggregation: the resulting multivariate time series exhibit spatial and temporal redundancy. This redundancy is then exploited to build mathematical models, useful for detection and diagnosis of data quality faults and business-related phenomena.

Type of Seminar:
Public Seminar
Speaker:
Dr. Janusz Milek
Predict AG, CH-4153 Reinach, and Automatic Control Laboratory, ETH Zürich
Date/Time:
Dec 19, 2001   17:15
Location:

ETH Zentrum, Gloriastrasse 35, 8006 Zurich, Building ETZ, room E6
Contact Person:

File Download:

Request a copy of this publication.
Biographical Sketch:
Janusz Milek obtained the M.Sc. degree in electronics engineering from Warsaw Technical University, Poland (1986), and the Ph.D. degree in the area of recursive parameter estimation from the ETH, Zürich (1995). He belongs to the research staff of the Automatic Control Laboratory, ETH Zürich. His research interests comprise various applications of statistical modeling (diagnosis of industrial processes, machine industry; information quality monitoring and data mining, Predict AG). He is a co-author of the forthcoming edition of Handbuch der industriellen Messtechnik, Oldenbourg Verlag.