This preview shows pages 1–4. 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: STAT5044: Regression and ANOVA, Fall 2010 Exam 1 on Nov 3 Your Name: Please make sure to specify all of your notations in each problem GOOD LUCK! 1 Problem# 1. (1.a) Write down a simple linear regression model. Specify each term as random or fixed. Specify each term as unknown term or known term. Y i =random, known, x i =fixed, known, i =random, unknown, , 1 , 2 =unknown, fixed. (1.b) In the simple regression model, suppose the value of the predictor X is replaced by cX , where c is some non zero constant. How are , 1 , 2 , and R 2 , and the ttest of H : 1 = 0 affected by this change? Let x i , new = cx i . Then , new = , 1 , new = 1 c 1 , 2 new = 2 , R 2 new = R 2 , t new = t (1.c) Suppose each value of the response of Y is replaced by dY , for some d 6 = 0. How are , 1 , 2 , and R 2 , and the ttest of H : 1 = 0 affected by this change? Let x i , new = dx i , , new = d , 1 , new = d 1 , 2 = d 2 2 , R 2 new = R 2 , t new = t 2 Problem# 2. Consider the following model Y i = + 1 X i + i , i = 1 ,..., n where E ( i ) = 0, Var ( i ) = 2 (unknown), Cov ( i , j ) = 0. (2.a) I would like to predict conditional mean of Y given a x new and a new observa tion y * given a x new . First I estimated a point estimator + 1 x new . What estimator would you use , 1 , and 2 ? Why? I would like to use LSE (below) because they are unbiased and have minimum vari ance among all the unbiased estimators. = ( X X ) T X Y , X = [ 1 , x ] 2 = Y ( I H ) Y n 2 (2.b) Under the assumption of i N ( , 2 ) , what is the distribution of + 1 x new ? Why? Let X new = ( 1 , x new ) and = ( , 1 ) T . Then Y new = X new = X new ( X X ) T X Y which is a linear form of Y. Since Y is following normal distribution and any linear form of Y is also following normal distribution, Y new is following normal distribution with E ( Y new ) = X new and Var ( Y new ) = 2 X new ( X X ) 1 X new...
View
Full
Document
This note was uploaded on 01/02/2012 for the course STAT 5044` taught by Professor Staff during the Fall '11 term at Virginia Tech.
 Fall '11
 Staff
 Linear Regression

Click to edit the document details