Statistical Methods in Physiomics Research and Clinical Applications
Co-Authors: Michael Imhoff1, Roland Fried2, Ursula Gather2
1 Department for Medical Informatics, Biometrics and Epidemiology, Ruhr-University Bochum, D-44780 Bochum, Germany
2 Department of Statistics, University of Dortmund, D-44221 Dortmund, Germany
With the continuing advances of genomics and proteomics research, physiological monitoring, and clinical data management, increasingly large amounts of data are available that contain information about the disease state of the individual critically ill or injured patient. Clinicians and researchers are challenged by the classification of disease entities like sepsis on the basis of high dimensional but sparsely populated data sets, by the identification of true failure from multivariate time series, by the analysis of complex spatial-temporal relationships, or by the extraction of reliable physiological information from noisy signals. In this talk, we give an overview of statistical methods, including graphical models, robust signal extraction, and multivariate time series models, that can help meet these challenges. Current and future applications of these and other methods, their potentials and limitations as well as research opportunities will be discussed.
Supported, in part, by the German Research Foundation (Deutsche Forschungsgemeinschaft; SFB 475).