Scribd
Upload
6 views6 pages

Lecture 11

The document discusses descriptive methods in regression and correlation, focusing on the relationship between dependent and independent variables through linear regression. It provides examples of simple regression calculations, including the interpretation of intercept and slope, and demonstrates the use of scatter plots to analyze relationships between variables. Additionally, it covers the least squares criterion for fitting lines and presents calculations for regression coefficients using both standard and shortcut formulas.

Uploaded by

Hafsa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
6 views6 pages

Lecture 11

The document discusses descriptive methods in regression and correlation, focusing on the relationship between dependent and independent variables through linear regression. It provides examples of simple regression calculations, including the interpretation of intercept and slope, and demonstrates the use of scatter plots to analyze relationships between variables. Additionally, it covers the least squares criterion for fitting lines and presents calculations for regression coefficients using both standard and shortcut formulas.

Uploaded by

Hafsa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 6

Descriptive Methods in Regression and Correlation

Regression: Study of relationship between a dependent variable and one or more


independent variables.

Simple regression: Deals with one independent variable

Linear Regression Equation: y=β 0 + β 1 x + e

e.g., Simple regression: Quantity demanded = β 0 + β 1(Price)+e

Linear equation, intercept and slope and their interpretation (Weissp-646,


Example 14.3)

Where,

β 0=Intercept ∨regression coefficient


β 1=Slop∨per unit rate of change ∈dependent variable
e=Residual∨error =difference between the actual∧estimated values
i.e. Residual: actual y – predicted y i.e. e = y− ^y, Geometrically the total area of
the squares is minimum possible.

e2

e1

EXAMPLE:
We have two quantitative variables the age and price of the onion. The age is
measured in years and is the predictor (x) while the price in hundreds dollars is the
response or dependent variable (y) for the problem. The observed data values are
given in the below table:
Scatter plot, [get a (rough) idea of direction, strength, linearity and also potential
presence of outliers in the (x ,y) relationship.]

Comment on Scatter plot:

As the dots are trending downward with small deviation of trend therefore we can
identify a strong negative relationship between the two under study variables i.e.
age of onion in years and the price of onion in hundred dollars.
Weissp-649, Example 14.6 (Ex Weiss14.59, 14.60 p-658, make scatter plot only
and comment on direction, strength, linearity)
Least Square Criterion p-648: The best fitting line corresponds to the intercept (
β 0 ¿and slope (β ¿¿ 1)¿such that the sum of squared residuals is minimum possible.
[Note in many cases intercept has no meaningful interpretation since x = 0 is not in
the data set for which regression is estimated].

Regression calculations:

As, Linear Regression Equation: y=β 0 + β 1 x + e

Where,
S xy ∑ (x −x)( y − y)
β 1= = (Standard Formula)
S xx ∑ (x−x )2
Also,
n ∑ xy−∑ x ∑ y
β 1= (Shortcut Formula)
n ∑ x −¿ ¿ ¿
2

Also,
β 0= y −β1 x
Now,
1) Calculation of β by standard formula.
1

x=5.273
y=88 .64

Car Age (x) Price (y) (x-x ̅) (y-y ̅) (x-x ̅)(y-y ̅) (x-x ̅)2
1 5 85 -0.273 -3.636 0.992 0.074
2 4 103 -1.273 14.364 -18.281 1.620
3 6 70 0.727 -18.636 -13.554 0.529
4 5 82 -0.273 -6.636 1.810 0.074
5 5 89 -0.273 0.364 -0.099 0.074
6 5 98 -0.273 9.364 -2.554 0.074
7 6 66 0.727 -22.636 -16.463 0.529
8 6 95 0.727 6.364 4.628 0.529
9 2 169 -3.273 80.364 -263.008 10.711
10 7 70 1.727 -18.636 -32.190 2.983
11 7 48 1.727 -40.636 -70.190 2.983
Sums 58 975 0 0 -408.909 20.182

S xy ∑ (x −x)( y − y) −408.909
β 1= = = =−20.26
S xx ∑ (x−x )2 20.182
2) Calculation of β 1by short cut formula.

n ∑ xy−∑ x ∑ y
β 1=
n ∑ x −¿ ¿ ¿
2

You can observe that the value of β 1 computed from both method is same
therefore, you can use either method for the estimation of β 1

Now we will compute β 0, which is β 0= y −β1 x


Here, x=5.273 , y=88.64 and β =−20.26
1

Therefore, β =88.64−(−20.26 ) (5.273 )=195.74


0

Therefore, the required regression equation to estimate the price


of onion in hundred dollars with the help of its age in years is
Price of Onion=195.74−20.26 (Age of onion)
Interpretation of intercept and slope of the equation
 The price of Onion was/ will be 195.74 (hundred dollars) when the
age of onion was/ will zero (years)
 Every year (age) the price of onion is decreased by 20.26 (hundred
dollars)
Now we can plot the regression line between the independent variable
(x) and the estimated values of dependent variable ^y on the same scatter
plot. The distance of actual data points (in green) with the regression line
(in blue) represents the amount of error/ residual in the prediction of
each point.
Another example:
a) Find the regression equation for the data points.

Check that n = 9,
∑ x =20682, ∑ y =3487.1, ∑ xy =9254378 , ∑ x2 =57414186 , ∑ y 2=1590653,
x=2298 , y=387.4556
S xx =∑ x −¿ ¿ ¿
2

( ∑ x )( ∑ y ) ( 20682 ) (3487.1 )
S xy =∑ xy − =9254378− =1241022.2
n 9
b 1=0.1255 , b0=99.007

Estimated Regression Equation: ^y =99.008+0.1255 x

x = living area of house (square feet), y = selling price ($1000)

b) Graph the regression equation and the data points. (We can use any two x values and find
corresponding ^y values from the estimated equation and use these points to plot the line).
c) Describe the apparent relationship between the two variables under consideration.
There appears to be (i) positive (ii) linear (iii) moderate relationship

d) Interpret the slope of the regression line.


A one Sq foot increase in living area is associated with increase in selling price by 0.1255
thousand of dollar (i.e. by $125.5)
e) Identify the predictor and response variables. Predictor: Living area , Response: Selling
Price

You might also like