LECTURE 1: DATA TABLE
Ruben Zamar
Deapartment of Statistics
UBC
September 3, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 1
()
September 3, 2013
1 / 37
PART I
DATA AND BACKGROUND
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 1
()
September 3
Exercises & Reading - Nov 10 - 23
1. Read Chapters 5 and 6 in the Course Notes
2. Solve Problems 5.4, 5.5 in Chapter 5 and Problems 6.1, 6.3,6.4, 6.5 in
Chapter 6
3. Suppose that you have a single factor, called A, with two levels (low =
-1 and high = 1).
Exercises & Reading- Sept 29-Oct 5
1. Read Chapter 4 in the Course Notes
2. Solve Problems 4.1, 4.2 in Chapter 4 of the Course Notes
3. Small simulation study.
(a) Simulate - using R - samples of 1000 independent chi-square random
variables with ncp=0 and
Exercises & Reading- Sept 22-28
1. Read Chapter 3 in the Course Notes
2. Solve Problems 3.1, 3.4 and 3.5 in Chapter 3 of the Course Notes
3. Let D be the design matrix for the linear model
y =D + "
0
B
B
B
D=B
B
@
Recall that
1
1
1
.
.
.
x11
x21
x31
.
.
.
Exercises & Reading- Sept 15-21
1. Read Chapter 2 in the Course Notes
2. Solve Problems 2.1 and 2.2 in the Course Notes
3. Let X be the design matrix for a linear model
0
1 x11 x12
B 1 x21 x22
B
B
X = B 1 x31 x32
B . .
.
.
@ . .
. .
.
1 xn1 xn2
x1p
x2p
x3
LECTURE 13: REVIEW
Ruben Zamar
Department of Statistics
UBC
November 26, 2013
Ruben Zamar
Department of Statistics
UBC
LECTURE 13
()
November 26, 2013
1 / 15
ORTHOGONAL DESIGN (Review)
Recall the design matrix
D=
1
x1
x2
xp
Suppose that
0
B
B
B
D 0 D = di
LECTURE 11: BLOCKING 2k DESIGNS
Ruben Zamar
Deapartment of Statistics
UBC
November 20, 2013
Ruben Zamar
Deapartment of Statistics
LECTURE 11: BLOCKING 2 k DESIGNS
UBC
()
November 20, 2013
1 / 12
LARGE NUMBER OF RUNS IN 2^k DESIGNS
With k factors at two le
LECTURE 10: 2k FACTORIAL DESIGNS
Ruben Zamar
Deapartment of Statistics
UBC
November 13, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 10
()
November 13, 2013
1 / 12
2^k x r FACTORIAL DESIGN
We have k factors at two levels and budget for
2k
r runs
LECTURE 8: FACTORIAL EXPERIMENTS
Ruben Zamar
Deapartment of Statistics
UBC
October 30, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 8
()
October 30, 2013
1 / 32
MOTIVATING EXAMPLE
An idealized hip bone implant is made of two materials, A and B,
LECTURE 9: ORTHOGONAL DESIGNS
Ruben Zamar
Deapartment of Statistics
UBC
November 4, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 9
()
November 4, 2013
1 / 14
ORTHOGONAL DESIGN
Recall the design matrix
D=
1
x1
x2
xp
Suppose that
0
B
B
B
D 0 D = d
LECTURE 7: FACTORIAL EXPERIMENTS
Ruben Zamar
Deapartment of Statistics
UBC
October 16, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 7
()
October 16, 2013
1 / 14
MOTIVATING EXAMPLE
An idealized hip bone implant is made of two materials, A and B,
LECTURE 5: RANDOMIZED BLOCK DESIGN
Ruben Zamar
Deapartment of Statistics
UBC
October 7, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 5
()
October 7, 2013
1 / 19
PART I
MOTIVATING EXAMPLE
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 5
()
Oct
LECTURE 3: ANALYSIS OF VARIANCE
Ruben Zamar
Deapartment of Statistics
UBC
October 1, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 3
()
October 1, 2013
1 / 48
PART I
LS IN MATRIX NOTATION
(REVIEW)
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE
LECTURE 6: EXAMPLES OF RANDOMIZED BLOCK
DESIGN
Ruben Zamar
Deapartment of Statistics
UBC
October 9, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 6
()
October 9, 2013
1 / 25
Randomized Blocks Design. Statistical Model
yij
= + i + j + ij
t
i
b
=
LECTURE 2: LEAST SQUARES
Ruben Zamar
Deapartment of Statistics
UBC
September 9, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 2
()
September 9, 2013
1 / 26
PART I
SIGNAL PLUS ERROR
MODELS
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 2
()
Sep
LECTURE 4: Re-Parametrization
Ruben Zamar
Deapartment of Statistics
UBC
September 29, 2013
Ruben Zamar
Deapartment of Statistics
UBC
LECTURE 4
()
September 29, 2013
1 / 19
PART I
AFFINE TRANSFORMATIONS
OF THE REGRESSION MODEL
PARAMETERS
Ruben Zamar
Deapar
STATISTICS 404
MIDTERM EXAMINATION
OCTOBER 2007
There are three questions. To get full marks, you must solve all three of them. Questions are equally weighted.
You might want to read the entire exam paper before you begin. Solutions must exhibit a clear,