=========Apologies for crosspostings ===================
Today's seminar:
Monday 8 November 1999 - 4pm, Room 122, Pearson Building
Department of Statistical Science - University College London
Speaker : Qiwei Yao (Institute of Mathematics & Statistics - University of
Kent at Canterbury)
Title: Varying-Coefficient Linear Models with Unknown Indices
Abstract:
We explore a class of varying-coefficient linear models in
which the coefficients are functions of an index defined as a unknown
linear combination of regressors and/or other variables. We search for
the least squared approximation from this class to a unknown
high-dimensional regression function, which is implemented through a
newly proposed hybrid back-fitting algorithm. The core of the algorithm
is the alternative iteration between estimating the index through a
one-step scheme (Bickel, 1975) and estimating coefficients through a
one-dimensional local linear smoothing. The generalized cross-validation
method for choosing bandwidth is incorporated into the algorithm in an
efficient manner. The locally significant variables are selected in
terms of a combined use of $t$-statistic and Akaike information criterion.
We further extend the algorithm for the models with two indices.
Simulation shows that the proposed methodology has appreciable
flexibility to model complex multivariate nonlinear structure and is
practically feasible with average modern computers. The methods are
further illustrated through the Canadian mink-muskrat data in
1925-1994 and the pound/dollar exchange rates in 1974-1983.
Useful Web sites:
The programme of our statistical seminars is available at:
http://www.ucl.ac.uk/Stats/research/journals.html
Map with instructions on how to get to our department:
http://www.ucl.ac.uk/Stats/map.html
Regards,
Kostas Skouras
----------------------------------------------------------------------------
---------------
K. Skouras, Lecturer in Statistics, Department of Statistical Science,
University College London, Gower Street, London WC1E 6BT, United Kingdom.
Tel: 0171-4193652 Fax: 0171-7383 4703
e-mail: [log in to unmask]
----------------------------------------------------------------------------
---------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|