Date & Time: January 24, 2013 :: 16:00-17:00
Venue: Ramanujan Hall
Title:The Dempster-Shafer Theory of Belief Functions for Managing Uncertainties: An Introduction
Speaker: Prof. R. P. Srivastava The University of Kansas, Lawrence, Kansas.
Abstract: The main purpose of the presentation is to introduce the Dempster-Shafer (DS) theory of belief functions and illustrate its flexibility through various applications. The DS theory was made popular by Shafer during the 1970s through his book A Mathematical Theory of Evidence published by Princeton University Press in 1976. Basically, the DS theory is a generalization of the Bayesian formalism; Bayes rule is a special case of [WINDOWS-1252?]Dempster’s rule. The Bayesian formalism makes direct probability statements about questions of interest, whereas the DS theory brings the probability statement on questions of interest in an indirect way. The advantage of the DS theory over Bayesian framework is in its ability to model ignorance explicitly. Evidence faced in the real world such as medicine, legal domain, or business, inherently contains partial ignorance. Thus, the DS theory is more appropriate for modeling and managing uncertainties in such situations. In fact, currently, the DS theory is being used in identifying objects and images in the real world. Also, it is being used for developing expert systems and for assessing risks such as fraud risk, information security risk, etc. The presentation will highlight some of these applicat