> Dear Colleagues,
> We continue our
> Colloquium
>
Qin Shao

Department of Statistics, The University of Georgia

January 30, Wednesday, Room 207, 2:00 pm.

Parsimonious Periodic Time Series Modeling

Abstract. Time series with periodic means and autocovariances frequently
arise in climatology, economics, hydrology, electrical engineering, etc.
Modeling methods for periodic series focus on the class of periodic
autoregressive moving-average (PARMA) models as the PARMA second moment
structure is indeed periodic. A ubiquitous problem in fitting a PARMA model
to a periodic series, however, lies with parsimony. Even very simple PARMA
models can have an inordinately large number of parameters--typically far
more than is necessary for an adequate statistical description of the
series. This talk presents results aimed at parsimonious PARMA model
development through the analysis of a daily temperature series from Griffin,
Georgia.

Qin Shao is a candidate for a tenure track position at our Department.