STA303H1-S: Methods of Data Analysis II Assignment 1 - Question 1
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
STA303H1-S: Methods of Data Analysis II
Assignment 1 - Question 1
Winter 2023
Q1 (6 points - each part 3 points):
1.) Consider the multiple linear regression model Y = Xβ + ε and the least squares estimate = (X\ X)−1 X\ Y . Show that
= β + Rε ,
where R = (X\ X)−1 X\ .
2.) Consider the multiple linear regression Y = Xβ + ε, where X is n × (p + 1) design matrix of rank p+1 < n, β = (β0 ,β1 , ··· ,βp )\ , and ε ∼ Nn (0,σ2 In ). Suppose we are interested in testing the hypothesis H0 : Cβ = 0, where C is a known q×(p+1) coefficient matrix of rank q ≤ p+1. This is known as the general linear hypothesis. The alternative hypothesis is HA : Cβ 0. Show that, if the null hypothesis is true, then the full-reduced-model test statistic is given by
(Cβˆ)\ [C(X\X)−1C\]−1 (Cβˆ)/q
RSS/(n − p − 1)
Hint:
❼ To understand the math, consider for example the model Y = β0 +β1X1 +β2X2 +β3X3 +ε
where we need to test H0 : 2β1 = 2β2 = β3 . This will be equivalent to test the hypothesis
H0 : Cβ = 0 where C = ).
❼ The minimization problem under the linear restriction can be solved by using the method
of the Lagrange multiplier, where you need to minimize (Y − Xβ)\ (Y − Xβ) + 2λCβ . Here λ denotes the Lagrange multiplier.
2023-02-04