Advanced Mathematics For Business: Topic 7: Regression Analysis 1

Advanced Mathematics for
Business
Topic 7:
Regression Analysis 1
V1.0
NCC Education Limited
Regression Analysis 1 Topic 7 - 1.2
Scope and Coverage

This topic will cover:
Simple linear regression
Pearsons (product moment) correlation coefficient
Spearmans (rank order) correlation coefficient
V1.0
Learning Outcomes
By the end of this topic students will be able to:
Understand a straight line fit to bivariate data
Calculate and interpret Pearsons correlation
coefficient
Calculate and interpret Spearmans correlation
coefficient
V1.0
Motivation
The search for, and strength of, predictors
-
V1.0
What is a good predictor of future job performance?

What is this products price-demand curve?
Which process factors affect production yield?
How does a particular share price move with the
market average?
savings y
The Linear Relationship

y = mx + c
income x
Assuming interval or ratio data.

V1.0
Least SSE Regression Criterion - 1

y
e5
e3
e4
e2
e1
V1.0

y
e5
e3
e4
e2
e1
V1.0
V1.0
Example - 1
A company has been carrying out experiments with
the position of a button on its website.
V1.0
Page
Position
(% vertical)
Click
Throughs
(%)
25
3.07
50
5.64
75
9.63
100
10.26
Example - 2
x
total
x2
25
3.07
76.75
625
50
5.64
282.00
2500
75
9.63
722.25
5625
100 10.26
1026.00
10000
2107
18750
250
28.6
mean 62.5
7.15
V1.0
xy
Example - 3
V1.0
Page
Position
(% vertical)
Click
Throughs
(%)
25
3.07
50
5.64
75
9.63
100
10.26
Exercise - 1
x
xy
x2
25
1.44
50
5.58
2500
75 14.64
5625
100
total
250
6.94
10000
18750
mean 62.5
V1.0
Exercise - 2
x
xy
x2
25
1.44
36
50
5.58
279
2500
75 14.64
1098
5625
100
6.94
694
10000
250
28.6
2107
18750
mean 62.5
7.15
total
V1.0
How Well Does the Line Fit? - 1

x
25
3.07
50
5.64
75
9.63
100 10.26
25
1.44
50
5.58
75 14.64
100
V1.0
6.94
unexplained
explained
V1.0
Pearson Correlation
V1.0
V1.0
xy
x2
y2
25
3.07
76.75
625
9.425
50
5.64
282.00
2500
31.810
75
9.63
722.25
5625
92.737
100
10.26
1026.00
10000 105.268
250
28.6
2107
18750 239.239

x
25
3.07
50
5.64
75
9.63
100 10.26
25
1.44
50
5.58
75 14.64
100
V1.0
6.94
Language of Correlation - 1
Sign
- r > 0 positive linear relationship
- r = 0 no linear relationship
- r < 0 negative linear relationship
V1.0
Language of Correlation - 2
Strength
- Physical sciences / engineering R2 > 0.6 often found
- Social sciences / policy R2 > 0.25 sometimes useful
- Business and management includes science & social
science!
- But you will see language like R2 > 0.8 strong, R2 > 0.5
moderate, R2 > 0.25 weak relationship
- Be careful; context and numbers often more informative
than descriptive word, but words help to communicate.
V1.0
What is r?
A
r=1
r = 0.9
D
r=0
V1.0
r = -0.7
r = 0.7
F
r = -0.9
Interpolation and Extrapolation

Interpolation - Estimates between values already known
Extrapolation - Estimates outside known values
V1.0
Basics of Simple Linear Regression

Plot scatter graph to intuit whether straight line is
reasonable
Look at r2 = R2 for strength of relationship
Look at sign of r for direction
- Check agrees with graph
Look at m for gradient of relationship

Use straight line equation to interpolate (with care)
Use straight line equation to extrapolate (with caution)
V1.0
Spearmans Rank Correlation
V1.0
Spearmans Correlation Example - 1

Two interviewers individually rank prospective job
candidates. What is the Spearman correlation
coefficient?
Candidate Interviewer 1 Interviewer 2
V1.0
Hidayat
Elisa
Nouman
Bernie
Li Ren
Ahere
Spearmans Correlation Example - 2
V1.0
Spearman Correlation - Ties?

For tied ranks use mean rank, then
Use formula for Pearson correlation
Candidate Interviewer 1 Interviewer 2
V1.0
Hidayat
OK
4
Elisa
Excellent
1
Nouman
Good
2
Bernie
OK
4
Li Ren
OK
4
Ahere
Poor
6
Recap
By the end of this topic students will be able to:
Understand a straight line fit to bivariate data
Calculate and interpret Pearsons correlation
coefficient
Calculate and interpret Spearmans correlation
coefficient
V1.0
Bibliography
Burton, G., Carrol, G. and Wall, S. Quantitative
Methods for Business and Economics. Longman.
Buglear, J. Quantitative Methods for Business.
Elsevier Butterworth Heinemann
Hinton, PR. Statistics Explained. Routledge
V1.0
Topic 7 Regression Analysis 1

Any Questions?
V1.0

Advanced Mathematics For Business: Topic 7: Regression Analysis 1

Uploaded by

Copyright:

Available Formats

Advanced Mathematics For Business: Topic 7: Regression Analysis 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advanced Mathematics For Business: Topic 7: Regression Analysis 1

Uploaded by

Copyright:

Available Formats

Advanced Mathematics for

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.2

Scope and Coverage

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.3

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.4

What is a good predictor of future job performance?

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.5

The Linear Relationship

Assuming interval or ratio data.

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.6

Least SSE Regression Criterion - 1

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.7

Least SSE Regression Criterion - 2

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.8

Least SSE Regression Criterion - 3

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.9

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.10

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.11

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.12

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.13

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.14

How Well Does the Line Fit? - 1

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.15

How Well Does the Line Fit? - 2

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.16

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.17

How Well Does the Line Fit? - 3

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.18

How Well Does the Line Fit? - 4

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.19

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.20

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.21

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.22

Interpolation and Extrapolation

NCC Education Limited

Regression Analysis 1 Topic 7 - 1.23

Basics of Simple Linear Regression

Look at m for gradient of relationship