Scribd
Upload
15 views2 pages

Test 2 2017

The document contains a test for a statistics course with 3 questions. Question 1 involves a linear model to measure the effect of chemicals and application methods on corn yield. Question 2 involves properties of linear models when the design matrix is less than full rank. Question 3 defines new random variables from existing random variables and calculates their mean, variance, and covariance.

Uploaded by

lindokuhlefezile76
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
15 views2 pages

Test 2 2017

The document contains a test for a statistics course with 3 questions. Question 1 involves a linear model to measure the effect of chemicals and application methods on corn yield. Question 2 involves properties of linear models when the design matrix is less than full rank. Question 3 defines new random variables from existing random variables and calculates their mean, variance, and covariance.

Uploaded by

lindokuhlefezile76
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

University of Kwazulu Natal Pietermartzburg campus

School of Mathematics , Statistics and Computer Science


Stat 301 Linear Models
Test 2 May 9, 2017
Instruction: Attempt all questions: Write neatly! Unreadable answers may not be
marked.
Time allowed : 2 hours.

1. An investigator wants to measure the effect on yield of a certain variety of corn of two
different chemical compounds and four different methods of applying these compounds.
Suppose  i is the effect of the i th chemical ( i= 1, 2) and  j is the effect of the j th
method of applying the compounds( j=1, 2, 3, 4). The investigator also assumes that the
effects are additive. The linear model for this experiment is

Yij =  + i +  j +  ij i = 1, 2 , and j = 1, 2, 3, 4

,
Find
a) Write down the design matrix (X)
b) Write the above linear model in matrix form.
c) How many parameters does the  vector has?
d) Write down the  vector.
[12]
2. Consider the linear model, y = X +  where  ~ N (0,  I ) and X is less
2

than full column rank.

a) Prove that the predicted values are unique


[3]
b) If q  is estimable, show that q b is the linear and unbiased estimator) of
q  where b = GX  Y for a G a generalized inverse of X X .
[8]

3. Let Z 1 , Z 2 , Z 3 be random variables with mean vector and covariance matrix

1  3 2 1
 = 2 ; V = 2 2 1
3 1 1 1

Define the new variables:


y1 = z1 + 2 z 3

y 2 = z1 + z 2 − z 3

y3 = 2 z1 + z 2 + z3

a) Find the mean vector and variance covariance matrix of ( y1 , y 2 , y3 ).


[4]
b) Find the mean and variance of
( y1 + y 2 + y3 )
3
[6]
c) Suppose w1 = y1 + y2 and w2 = y1 − y2 . Find the mean vector and variance
covariance matrix of (w1, w2 ) .
[7]

You might also like