Analysis of Variance and Design of Experiments
Analysis of Variance and Design of Experiments
Analysis of Variance and Design of Experiments
of Variance and Design of
Experiments
General Linear Hypothesis and Analysis of Variance
:::
Lecture 4
ANOVA models and Least Squares Estimation of
Parameters
Shalabh
Department of Mathematics and Statistics
Indian Institute of Technology Kanpur
Slides can be downloaded from http://home.iitk.ac.in/~shalab/sp1
Regression model for the general linear hypothesis:
Let Y1 , Y2 ,..., Yn be a sequence of n independent random variables
associated with responses. Then we can write it as
p
Yi j xij i , i 1, 2,..., n; j 1, 2,..., p
j 1
where 1 , 2 ,..., p are the unknown parameters and xij’s are the
known values of independent covariates X 1 , X 2 ,..., X p , i ’s are
i.i.d. random error component with
E ( i ) 0 , V ar ( i ) 2 , C ov ( i , j ) 0( i j ).
2
Relationship between the regression and ANOVA models:
Now consider the case of one‐way model and try to understand
its interpretation in terms of the multiple regression model.
3
Relationship between the regression and ANOVA models:
The linear model now can be expressed as
Y 0 1 X 1 2 X 2 ... p X p
by defining
and so on.
5
Relationship between the regression and ANOVA models:
If the experiment with 1 Kg. of fertilizer is repeated n1 number of
times then n1 observation on response variables are recorded
which can be represented as
6
Relationship between the regression and ANOVA models:
If X2 = 1 is repeated n2 times, then on the same lines n2
number of times then n2 observation on response variables are
recorded which can be represented as
7
Relationship between the regression and ANOVA models:
The experiment is continued and if Xp = 1 is repeated np times,
then on the same lines
Yp1 0 1.0 2 .0 ... p .1 P1
Yp 2 0 1.0 2 .0 ... p .1 P 2
Ypn p 0 1.0 2 .0 ... p .1 pn p
8
Relationship between the regression and ANOVA models:
All these n1 , n 2 ,.., n p observations can be represented as
y
11
1 1 0 0 0 0 11
y12 12
1 1 0 0 0 0
y1n 1n
1 1 0 0 0 0
1
1
y21 1 0 1 00 0 21
y
0
1 0 1 0 0 0 22
22
1
y
2 n2 1 0 1 0 0 0
2 n2
p
y p1 1 0 0 00 1 p1
1 0 0 00 1
p2
y p2
1 0 0 00 1
y pn pn
p p
or
Y X . 9
Relationship between the regression and ANOVA models:
In the two‐way analysis of variance model, there are two
covariates and the linear model is expressible as
Y 0 + 1 X 1 2 X 2 ... p X p 1 Z 1 2 Z 2 ... q Z q
X 1 , X 2 ,..., X p
where denotes, e.g., the p levels of the quantity of
Z 1 , Z 2 ,..., Z q
fertilizer, say 1 Kg., 2 Kg.,...,p Kg. and denotes, e.g.,
the q levels of level of irrigation, say 10 Cms., 20 Cms.,…100 Cms.
etc. The levels are the counter variable
X 1 , X 2 ,..., X p Z 1 , Z 2 ,..., Z q
indicating the presence or absence of the effect as in the earlier
case.
10
Relationship between the regression and ANOVA models:
If the effect of X1 and Z1 are present, i.e., 1 Kg of fertilizer and 10
Cms. of irrigation is used then the linear model is written as
Y 0 1.1 2 .0 ... p .0 1.1 2 .0 ... p .0
0 1 1 .
13
Relationship between the regression and ANOVA models:
14
Analysis of variance:
Analysis of variance is a body of statistical methods of analyzing
the measurements assumed to be structured as
16
Least squares estimate of
Let y1 , y2 ,..., yn be a sample of observations on Y1, Y2 ,..., Yn .
The least‐squares estimate of is the values ˆ of for which the
sum of squares due to errors, i.e.,
n
S 2 i2 ' ( y X )( y X )
i 1
yy 2 X ' y X X
17
Least squares estimate of
Differentiating S2 with respect to and substituting it to be
zero, the normal equations are obtained as
dS 2
2 X X 2 X y 0
d
or X X X y
ˆ ( X X ) 1 X y
18
Least squares estimate of
ˆ (LSE) is the best linear unbiased estimator of in the sense of
having minimum variance in the class of linear and unbiased
estimator.
19
Least squares estimate of
A question arises that what are the conditions under which a
linear parametric function L admits a unique least‐squares
estimate in the general case.
20