Prof. Henry P. Wynn, Ph.D.
Two rather different uses of monomial ideals have been developed in probability and statistics. The first is related to discrete tube theory which gives tight Bonferroni-type bounds for use in areas such as in reliability theory, network theory and test situations such as scan statistics. This theory applies to subsets of the appropriate probability space. The second theory applies to random variables and is used to capture model assumptions such as conditional independence in hierarchical models and interactions in polynomial regression models. Both use the theory of minimal free resolution, Betti numbers, Alexander duality etc. The talk will discuss situations in which both theories may be used and whether, in some sense, the theories are dual.
Datum: 10.01.2013, Raum: G03-106, Zeit: 17:00