This preview shows pages 1–5. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Lecture 19: Robust & Quantile regression Stat 704: Data Analysis I, Fall 2010 Tim Hanson, Ph.D. University of South Carolina T. Hanson (USC) Stat 704: Data Analysis I, Fall 2010 1 / 17 Chapter 11 11.3 Influential cases rem. measure: Robust regression 11.3: Robust regression * Leverages h ii and deleted residuals t i useful for finding outlying x i and Y i (w.r.t. the model) cases. * Cook’s D i and DFFIT i indicate which cases are highly influencing the fit of the model, i.e. the OLS b . * What to do with influential and/or outlying cases? Are they transcription errors or somehow unrepresentative of the target population? * Outliers are often interesting in their own right and can help guide the building of a better model. * Robust regression dampens the effect of outlying cases on estimation to provide a better fit to the majority of cases. * Useful in situations when there’s no time for “influence diagnostics” or a more careful analysis. T. Hanson (USC) Stat 704: Data Analysis I, Fall 2010 2 / 17 Chapter 11 11.3 Influential cases rem. measure: Robust regression * Robust regression is effective when the error distribution is not normal, but heavytailed. * Mestimation is a general class of estimation methods. Choose β to minimize Q ( β ) = n X i =1 ρ ( Y i x i β ) , where ρ ( · ) is some function. * ρ ( u ) = u 2 gives OLS b . * ρ ( u ) =  u  gives L 1 regression, or least absolute residual (LAR) regression. * Huber’s method – described next – builds a ρ ( · ) that is a compromise between OLS and LAR. T. Hanson (USC) Stat 704: Data Analysis I, Fall 2010 3 / 17 Chapter 11 11.3 Influential cases rem. measure: Robust regression Iteratively Reweighted Least Squares Outlying values of r j i = y i x i b j are (iteratively) given less weight in the estimation process. (Starting b ) OLS: w i = 1 / e 2 i from e = Y bX . Set j = 1....
View
Full
Document
This note was uploaded on 12/14/2011 for the course STAT 704 taught by Professor Staff during the Fall '11 term at South Carolina.
 Fall '11
 Staff

Click to edit the document details