Upload
doreen-bryant
View
213
Download
1
Embed Size (px)
Citation preview
1
Introduction
What does it mean when there is a strong positive correlation between x and y ?
Regression analysis aims to find a precise formula to relate the movements of y to those of x
The use of regression requires a good deal of thought and a good dose of skepticism
2
Example: Sales vs. Advertising
Advertising (mil.$) Sales (mil.$)1.2 1201.6 1901.8 2602.2 2602.6 3003.1 2903.4 3303.6 3304.0 3404.2 310
3
Example: Sales vs. Advertising
0
50
100
150
200
250
300
350
400
0 1 2 3 4 5
4
Example: Sales vs. Advertising
It is believed that sales, S, are tied to advertising, A, by a simple linear equation:
What do 0 and 1 represent?
How can we find 0 and 1 ?
AS 10
5
Sales vs. Advertising
The linear relation is usually not exact.
A more realistic model:
Where 0 and 1 are regression coefficients.
ii10i AS
6
Measurement Error
Properties of distribution A mean of zero Symmetry around zero An assignment of greater probability to
small errors than to larger ones Errors are assumed to be:
Independent Have same variance
(homoscedasticity)
i
7
Method of Least Squares
Consider the simple formula:
Where the measurement errors are independent samples from N(0, )
How to find the estimators of 0 and 1?
xY 10
8
Choosing the best line
y
x
I II
III
9
Least Squares Estimates
Suggest an index to measure discrepancy between points and line
Focus on vertical disparities between points and line
Sum of the square of the deviations:
n
1i
n
1i
2i10i
2i10 )xy(),(L
10
Least Squares Estimates
n1i i
n1i i
n
1i
2n
1ii
2i
n
1ii
n
1iin
1iii
1
10
x)n/1(x and y)n/1(y where
n
x
x
n
xy
xyˆ
xˆyˆ