Week of October 11 - 15

Tuesday, October 12

 2:30 ALGEBRA SEMINAR  - Room 447 MSC
      Steve McCleary, Mathematics and Statistics, BGSU
      "Lattice-ordered permutation groups, Part IV"
      This is the fourth in a series of talks.

 2:30 SCIENTIFIC COMPUTATION SEMINAR  - Room 459 MSC
      Mark Jarvis, Mathematics and Statistics, BGSU 
      "Fourier domain processing of real-time signals"

      Abstract: When applying Fourier Methods to even well defined
        sampled functions in L^2, there are still effects such as
        Gibbs' phenomena which prevent accurate reconstruction of the
        signal.  When discussing real-time signals, a time dependent
        frequency spectrum must also be accounted for.  Though there
        are analog methods for processing such signals, by far most
        methods are numerical.

        A graphical discussion of linear systems' response will
        introduce issues and effects of the Analog to Digital,
        Windowing, and Digital to Analog processes on real-time signal
        spectrums such as those obtained in voice and image
        processing.

        By understanding the limitations of Fourier Processing, we
        hopefully will better understand the reasons behind the highly
        visible efforts in Wavelet Transform research.

Wednesday, October 13

 2:30 ANALYSIS SEMINAR  - Room 459 MSC
      Ron Taylor, Mathematics and Statistics, BGSU 
      "A Banach space operator with a prescribed orbit"

Thursday, October 14

 1:30 LUKACS LECTURE  - Room 400 MSC
      Raju Govindarajulu, Distinguished Lukacs Professor, BGSU and
                          University of Kentucky
      "Application of rank tests to random effects model"

Friday, October 15

 1:30 LUKACS LECTURE  - Room 400 MSC
      Raju Govindarajulu, Distinguished Lukacs Professor, BGSU and
                          University of Kentucky
      "Chernoff-Savage class of statistics: asymptotic theory"
      This will be a two-hour talk.

 3:30 Refreshments
 3:45 COLLOQUIUM  - Room 459 MSC
      Truc Nguyen, Mathematics and Statistics, BGSU
      "Exact EDF goodness-of-fit tests for inverse Gaussian distributions"
      Abstract: Characterizations of inverse Gaussian distributions in
        different cases of unknown parameters based on the uniformly
        minimal variance unbiased estimator (UMVUE) of the density
        function are studied.  Using these characterization results as
        transformations to change the composite null hypothesis that
        "F is an inverse Gaussian distribution" to an equivalent
        simple null hypothesis, then exact empirical distribution
        function (EDF) goodness-of-fit tests for inverse Gaussian
        distributions are constructed.  In the case of an inverse
        Gaussian(m,b) distribution with known b and unknown m, a
        chi-square test is also proposed.  The powers of these tests
        are estimated by Monte-Carlo method at several different
        alternative distributions.