106
reads

Tzee-Ming Huang

2013-07-02
13:30:00 - 14:20:00

101 , Mathematics Research Center Building (ori. New Math. Bldg.)

It was demonstrated that BARS, a free-knot spline fitting method in a Bayesian framework, outperformed SARS, a free-knot spline fitting method based on penalized likelihood, in various examples. Since BARS estimator (the posterior mean) is an average of posterior curves, it is of interest to know that whether using a frequentist model average approach can give a fitting method that is as competitive as BARS. In this talk, a fitting method based on frequentist model average will be introduced and comparative results will be presented.