previous | colloquium schedule | next

Colloquium - 21 Oct 1999

Department of Statistics
Texas A&M University

ABSTRACT: An omnipresent problem in statistics is testing whether or not some model for a function is in agreement with observed data. In this talk, various nonparametric tests based on smoothing ideas and/or model selection methods are described and illustrated. Data from plant-breeding experiments are used to illustrate a smoother-based test of the equality of two probability distributions.

Tests based on data-driven model selectors are considered in cases where the observed likelihood depends upon a function,g, and a vector, eta, of nuisance parameters. Examples of such situations abound in linear models, generalized linear models, time series, and many other areas. Assuming that the likelihood function is correctly specified up to the form of g, our interest is in testing the fit of a parametric model for g. Aerts, Claeskens and Hart (1999) proposed methodology for this testing problem. We describe and illustrate their methods, whose distinguishing feature is the use of a data-driven choice of model dimension.

Refreshments will be served in the Blocker Building, Room 447, at 3:30 p.m.

Copyright © Unless otherwise stated, all contents of these web pages are copyright by their respective authors or by the Department of Statistics, Texas A&M University, 1995-1999. Comments and questions about this site should be sent to web@stat.tamu.edu.