Weekly Calendar of Seminars, Talks, and Events Department of Mathematics & Statistics Bowling Green State University Week of February 1-5, 2010 Monday, February 1, 2010 2:30 PM Analysis Seminar 459 MSC Melanie Henthorn, BGSU The growth of eigenvalues and its relationship to spectral synthesis, Part 2 Tuesday, February 2, 2010 10:45 AM Math 1220 Instructor Meeting 400 MSC Wednesday, February 3, 2010 11:30 AM Statistics Seminar 459 MSC Adam Combs, BGSU A look at formulating prior distributions for analysis in clinical trials Turning informally expressed opinions into a mathematical prior distribution is a very difficult, and esoteric, aspect of Bayesian analysis. An overview of five broad approaches for obtaining prior distributions will be given, as it pertains to doing Bayesian analysis in clinical trials. These approaches are: elicitation of subjective opinion, summarizing past evidence, use of default priors, robust priors and estimation of priors using hierarchical models. 3:30 PM Algebra Seminar 459 MSC Ben Otto, BGSU Finite Calculus 4:30 PM Graduate Committee Meeting 400 MSC Thursday, February 4, 2010 11:00 AM Math 1150 Instructor Meeting 400 MSC 2:30 PM Advisory Committee Meeting 400 MSC 4:00 PM Calculus Seminar 459 MSC Indeterminate forms and L'Hopital's Rule Mary Koshar session leader 7:00 PM The BGSU student Actuarial Science Society hosts Lonie Moore, Associate Towers-Watson 459 MSc For an informal discussion of the SOA/CAS actuarial exams, exam-prep strategies, study materials, etc. Lonie is a 2008 graduate of the BGSU actuarial science program. Please join us! All are welcome - refreshments provided. Friday, February 5, 2010 3:30 PM COLLOQUIUM 459 MSC Dr. Guohui Song, Illinois Institute of Technology Fast algorithms for kernel-based scattered data approximation Scattered data approximation deals with the problem of reconstructing an unknown function from given scattered data. It has applications in a variety of fields such as surface construction, the numerical solution of partial differential equations, statistical learning, and parameter estimation. A popular approach in scattered data approximation is the kernel-based regularization method that consists of calculating the inverse of a matrix generated by a kernel function and the given data. However, the computational cost of inverting this matrix is a major concern especially for high-dimensional data. We introduce some fast algorithms for calculating the inverse by approximating the kernel matrix with a related multilevel circulant matrix so that the fast Fourier transform can apply to reduce the computational cost. We also give a super fast algorithm for high-dimensional tensor-product kernels. An application in political science will be mentioned if time permits. This is joint work with Dr. Yuesheng Xu at Syracuse University.
A list of mathematics seminars by subject and other seminars at BGSU is available here.
If you have comments or material for the calendar, send e-mail to Anita Serda,
If you wish to be placed on the e-mail distribution list, send e-mail to Craig Zirbel,