[Statlist] reminder SfS

[Christina Kuenzli]

                       ETH and University of Zurich 
      Proff. A.D. Barbour -- P. Buehlmann -- F. Hampel -- H.R. Kuensch 

*************************************************************************
           We are pleased to announce the following seminar
*************************************************************************

  May, 16, 2003,  15.15 
  Leonhardstrasse 27, 8006 Zürich, LEO C15

  Precise and imprecise Bayesian probability networks for
  prediction, inference, and decision-making in environmental science 

  Mark Borsuk 
  Systemanalyse, Integrated Assessment und Modellierung, EAWAG, Dübendorf

  Scientists are often asked to support decisions by providing the
  probabilistic link between actions and outcomes.  However, this link
  generally represents a complex causal chain, crossing many scientific
  disciplines.  Therefore, it is useful to have a framework for decomposing
  the prediction problem into smaller parts that can each be addressed
  separately and then later assembled into an integrated model.  We have
  found that a class of causal models called probability networks is
  particularly useful for this purpose.  A probability network (also called
  a Bayesian, or belief, network) is the combination of a graphical
  depiction of the causal relationships among the most important variables
  in a system with a quantification of these relationships using
  conditional probabilities.  The graphical depiction facilitates the
  identification of conditional independencies that allow for model
  decomposition, while the conditional probabilities quantitatively
  describe the causal relationships accounting for uncertainties.  Once all
  parts of the network are fully specified, probabilistic predictions of
  model endpoints can be generated for any set of values for the marginal
  input variables.  Unfortunately, as with most other probabilistic models
  in current use, the construction of a probability network requires the
  specification of many precise probabilities describing the relationships
  among variables.  However, there is nearly always either a shortage of
  data, disagreement among experts, or limited time and resources for
  detailed analysis, resulting in some ambiguity in probability
  determinations.  Therefore, probabilities might be better represented as
  imprecise quantities described by bounded sets.  We are currently
  investigating opportunities and implications of this generalization for
  probability network models used for environmental decision-making.
  Concepts and ideas will be presented within the context of real-world
  environmental examples. 

*************************************************************************

  May 23, 2003,  15.15
  Leonhardstrasse 27, 8006 Zürich, LEO C15

  Hierarchical Testing Design for Pattern Recognition

  Donald Geman, Dept. of Mathematics Sciences, 
  Johns Hopkins University, Baltimore 
  
*************************************************************************
  June 6, 2003, 15.15
  Leonhardstrasse 27, 8006 Zürich, LEO C15

  Guilt by association: detecting human disease genes by analyzing DNA
  sequence patterns

  Anja Wille, Seminar für Statistik, ETH Zentrum

*************************************************************************
(LEO is close to the main building, across the hill-side station of the
'Polybahn') 
Overview maps of ETH : http://www.ethz.ch/search/orientation_en.asp

Further information: Christina Kuenzli, Statistics Seminar of ETH Zurich 
**************************************************************************

Everybody is kindly invited
   ________________  ___
  / ____/_   __/  /_/  /
 / __/_  /  / /  __   / 
/_____/ /__/ /__/ /__/ 
Eidgenoessische Technische Hochschule Zuerich
Swiss Federal Insitute of Technology Zurich
________________________________________________________
Christina Kuenzli            <kuenzli at stat.math.ethz.ch>
Seminar fuer Statistik      
Leonhardstr. 27,  LEO D11          phone: +41 1 632 3438         
ETH-Zentrum,                       fax  : +41 1 632 1228 
CH-8092 Zurich, Switzerland        http://stat.ethz.ch/~