This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Tutorial 6: suggested solutions 1. Suppose m ( x 1 , ..., x p ) = φ ( α 1 x 1 + ... + α p x p ). Prove that ∂m ( x 1 ,...,xp ) ∂x 1 ∂m ( x 1 ,...,xp ) ∂x 2 ... ∂m ( x 1 ,...,xp ) ∂xp = φ ( α 1 x 1 + ... + α p x p ) α 1 α 2 ... α p proof: Because ∂m ( x 1 , ..., x p ) ∂x k = φ ( α 1 x 1 + ... + α p x p ) α k 2. For a singleindex model Y = φ ( α > X ) + ε . Suppose ( X i , y i ) are the observations and the estimator for α is ˆ α . For a new point X , predict the expectation of response using local linear kernel smoothing. the prediction value is ˆ g (ˆ α > X ) = ∑ n i =1 { s n, 2 (ˆ α > X ) K h (ˆ α > ( X i X )) s n, 1 (ˆ α > X ) K h (ˆ α > ( X i X ))ˆ α > ( X i X ) /h } Y i s n, 2 (ˆ α > X ) s n, (ˆ α > X ) s 2 n, 1 (ˆ α > X ) where s n,k (ˆ α > X ) n X i =1 K h (ˆ α > ( X i X )) { ˆ α > ( X i X ) /h } k , k = 0 , 1 , 2 3. For the ozone concentration...
View
Full
Document
This note was uploaded on 10/04/2010 for the course STAT ST4240 taught by Professor Xiayingcun during the Fall '09 term at National University of Singapore.
 Fall '09
 XIAYingcun

Click to edit the document details