This preview shows pages 1–6. 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 DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Nonparametric regression using smoothing splines I Smoothing is fitting a smooth curve to data in a scatterplot I Will focus on two variables: Y and one X I Our model: y i = f ( x i ) + i , where 1 , 1 ,... n are independent with mean 0 I f is some unknown smooth function I Up to now f has a specified form with unknown parameters I f could be linear or nonlinear in the parameters, I functional form always specified I If f not determined by the subject matter, we may prefer to let the data suggest a function form c 2011 Dept. Statistics (Iowa State University) Stat 511 section 30 1 / 20 I Why estimate f ? I can see features of the relationship between X and Y that are obscured by error variation I summarizes the relationship between X and Y I provide a diagnostic for a presumed parametric form I Example: Diabetes data set in Hastie and Tibshiranis book Generalized Additive Models I Examine relationship between age of diagnosis of diabetes and log of the serum Cpeptide concentration I Heres what happens if we fit increasing orders of polynomial, then fit an estimated f c 2011 Dept. Statistics (Iowa State University) Stat 511 section 30 2 / 20 5 10 15 3.0 4.0 5.0 6.0 Age at Diagnosis log Cpeptide concentration c 2011 Dept. Statistics (Iowa State University) Stat 511 section 30 3 / 20 5 10 15 3.0 4.0 5.0 6.0 Age at Diagnosis log Cpeptide concentration linear fit c 2011 Dept. Statistics (Iowa State University) Stat 511 section 30 4 / 20...
View
Full
Document
 Spring '08
 Staff

Click to edit the document details