Unimodal Density Estimation with Application to Robust Regression
Mary Meyer ,Department of Statistics, Colorado State University
Monday,September 19, 2011
4:00 p.m., room 223, Weber Bldg
A smooth unimodal regression spline density estimator is described using a least-squares criterion. The estimator is easy compute and is shown to have an optimal convergence rate. This leads to a new method for robust regression, where the regression coefficients and the error density are estimated simultaneously. The method is shown to be similar to M-estimation.