Professional Documents
Culture Documents
Unit - V
Regression Analysis
Dr. E. Suresh,
Assistant Professor, Department of Mathematics,
SRM Institute of Science and Technology,
Kattankulathur - 603203.
Regression Analysis
Regression
Regression
Mathematical Relationship
Regression
Mathematical Relationship
Estimation
Regression
Mathematical Relationship
Estimation
Y = a + bX
Y = a + bX
1 X is indepedent variable
Y = a + bX
1 X is indepedent variable
2 Y is depedent variable
Y = a + bX
1 X is indepedent variable
2 Y is depedent variable
3 a is intercept
Y = a + bX
1 X is indepedent variable
2 Y is depedent variable
3 a is intercept
4 b is slope
Y = a + bX
1 X is indepedent variable
2 Y is depedent variable
3 a is intercept
4 b is slope
5 b is regression coefficient of Y on X . i.e., byx
Regression coefficient
P P P
n XY − ( X ) ( Y )
byx =
n X 2 − ( X )2
P P
X = a + bY
X = a + bY
1 Y is indepedent variable
X = a + bY
1 Y is indepedent variable
2 X is depedent variable
X = a + bY
1 Y is indepedent variable
2 X is depedent variable
3 a is intercept
X = a + bY
1 Y is indepedent variable
2 X is depedent variable
3 a is intercept
4 b is slope
5 b is regression coefficient of X on Y . i.e., bxy
σx
X −X =r Y −Y
σy
σx
X −X =r Y −Y
σy
σx
X −X =r Y −Y
σy
σx
X −X =r Y −Y
σy
X - X = bxy Y − Y
σx
X −X =r Y −Y
σy
X - X = bxy Y − Y
P P P
n XY − ( X ) ( Y )
Regression coefficient bxy =
n Y 2 − ( Y )2
P P
1 The regression lines were passes through the point X,Y
1 The regression lines were passes through the point X,Y
1 The regression lines were passes through the point X,Y
1 The regression lines were passes through the point X,Y
p
2 If bxy , byx are positive then r = + bxy × byx
p
2 If bxy , byx are positive then r = + bxy × byx
p
3 If bxy , byx are negative then r = − bxy × byx
Problem No. 1
Obtain the two regression lines and correlation coefficient from the
following data
X : 1 2 3 4 5 6 7 8 9
Y : 9 8 10 12 11 13 14 16 15
X Y X2 Y2 XY
1 9
2 8
3 10
4 12
5 11
6 13
7 14
8 16
9 15
X2 Y2
P P P P P
X Y XY
Here
Here
P P P
n XY − ( X ) ( Y )
byx =
n X 2 − ( X )2
P P
Here
P P P
n XY − ( X ) ( Y ) 9(597) − (45)(108)
byx = P 2 = = 0.95
9(285) − (45)2
P 2
n X − ( X)
Here
P P P
n XY − ( X ) ( Y ) 9(597) − (45)(108)
byx = P 2 = = 0.95
9(285) − (45)2
P 2
n X − ( X)
Y − 12 = (0.95)(X − 5)
Here
P P P
n XY − ( X ) ( Y ) 9(597) − (45)(108)
byx = P 2 = = 0.95
9(285) − (45)2
P 2
n X − ( X)
Y − 12 = (0.95)(X − 5)
Simplifying we get
Y = 0.95X + 7.25
Here
Here
P P P
n XY − ( X ) ( Y )
bxy =
n Y 2 − ( Y )2
P P
Here
P P P
n XY − ( X ) ( Y ) 9(597) − (45)(108)
bxy = P 2 = = 0.95
9(1356) − (108)2
P 2
n Y −( Y)
Here
P P P
n XY − ( X ) ( Y ) 9(597) − (45)(108)
bxy = P 2 = = 0.95
9(1356) − (108)2
P 2
n Y −( Y)
X − 5 = 0.95(Y − 12)
Here
P P P
n XY − ( X ) ( Y ) 9(597) − (45)(108)
bxy = P 2 = = 0.95
9(1356) − (108)2
P 2
n Y −( Y)
X − 5 = 0.95(Y − 12)
Simplifying we get
X = 0.95Y − 6.4
p
r = ± byx × bxy
p
r = ± byx × bxy
p
r= byx × bxy
p
r = ± byx × bxy
p √
r= byx × bxy = 0.95 × 0.95
p
r = ± byx × bxy
p √
r= byx × bxy = 0.95 × 0.95 = 0.95
When X = 6.2
Y = 0.95X + 7.25 = (0.95 × 6.2) + 7.25 = 13.14.
Dr. E. Suresh 18MAB303T - BioStatistics for Biotechnologists
Simple Shortcut
coefficient of y
bxy = −
coefficient of x
coefficient of y
bxy = −
coefficient of x
r 2 = bxy · byx ≤ 1
Dr. E. Suresh 18MAB303T - BioStatistics for Biotechnologists
Simple Shortcut
coefficient of y
bxy = −
coefficient of x
r 2 = bxy · byx ≤ 1
Dr. E. Suresh 18MAB303T - BioStatistics for Biotechnologists
Problem No. 3
Problem No. 3
In a partially destroyed laboratory record of an analysis of
correlation data, following result only legible. Variance of X = 9.
The regression equations are 8X − 10Y + 66 = 0 and
40X − 18Y = 214. What are
(i) The means value of X and Y
(ii) the correlation coefficient bet X and Y
(iii) the S.D of Y .
Solution:
Since the regression lines were passes through the point X , Y .
Hence equations (1) and (2) becomes
Since the regression lines were passes through the point X , Y .
Hence equations (1) and (2) becomes
Since the regression lines were passes through the point X , Y .
Hence equations (1) and (2) becomes
Since the regression lines were passes through the point X , Y .
Hence equations (1) and (2) becomes
Since the regression lines were passes through the point X , Y .
Hence equations (1) and (2) becomes
X = 13, Y = 17
⇒ 10y = 8x + 66
⇒ 10y = 8x + 66
8 66
⇒y = x+
10 10
y = 0.8x + 6.6
8
byx = = 0.8
10
Problem No. 3
(ii) To find the correlation coefficient bet X and Y
Let us consider equation (1)
⇒ 10y = 8x + 66
8 66
⇒y = x+
10 10
y = 0.8x + 6.6
⇒ 10y = 8x + 66
8 66
⇒y = x+
10 10
y = 0.8x + 6.6
⇒ 10y = 8x + 66
8 66
⇒y = x+
10 10
y = 0.8x + 6.6
8
byx = = 0.8
10
18 214
⇒x = y+
40 40
x = 0.45x + 5.35
18 214
⇒x = y+
40 40
x = 0.45x + 5.35
The regression coefficient of x on y is
18 214
⇒x = y+
40 40
x = 0.45x + 5.35
The regression coefficient of x on y is
18
byx = = 0.45
40
The correlation coefficient between X and Y is
18 214
⇒x = y+
40 40
x = 0.45x + 5.35
The regression coefficient of x on y is
18
byx = = 0.45
40
The correlation coefficient between X and Y is
p √
r = byx × bxy = 0.8 × 0.45 = 0.60.
Problem No. 3
(2) can be written as
(2) ⇒ 40X − 18Y = 214
18 214
⇒x = y+
40 40
x = 0.45x + 5.35
The regression coefficient of x on y is
18
byx = = 0.45
40
The correlation coefficient between X and Y is
p √
r = byx × bxy = 0.8 × 0.45 = 0.60.
Dr. E. Suresh 18MAB303T - BioStatistics for Biotechnologists
Problem No. 3
8 r σy 8
byx = ⇒ =
10 σx 10
r σy 8 0.6 × σy
⇒ = ⇒ = 0.8
σx 10 3
0.8 × 3
⇒ σy = = 4 ⇒ σy = 4
0.6