Session 3,4

To identify min no.
segments that explain max variance in data

Identify min no. of segments that will cater to max no. of consumer
First thing we do is
1. Put everything in one folder
2.Set working directory to source file location
3. Import data
4.Start running codes
Bartlett test the hypothesis if there is any correlation

Identity matrix is significantly different from identity matrix
KMO value
MSA is a measure of sampling adequacy
Factor matrix
the number inside are called factor loadings; simple correlation of a particular variable with the
respective factor
H2: the amount of variance being explained by total no. of factors

U2: is uniqueness
Relevant factors
(Eigen value) SS loading is sum of sq. loadings of previous vector column
All those factors or components or segments which has eigen value (ss loadings) greater than 1.
If SS loadings is greater than one significant variance in data is explained by that component
Those are the number of factors or segments we consider
That means we will group the component in two sectors
There are only 2 components 93% OF VARIANCE in variable 1 is explained by two compenenst
pca2
KMO: Square of all the Correlation matrix vector. Column wise-sum (downwards)
Anti image correlation matoci
KMO= N/(N+Na)
Which variable will fall in which component?
The threshold value of factor loadings should be 0.5

The threshold value of KMO should be 0.6
We should minimize the cross loading

theoreitically there should not be cross loading but practically its not possible
In simpler terms, varimax rotation reorganizes the components to make them easier to understand.
It tries to make sure that each component is represented by a few variables with high loadings,
making it clearer which variables are most important for each component.
Here improvement in results means minimization of cross loading

We remove the outliers to remove the cross loadings
We identifies 2 segments
V1,V3,V5: S1: Heath consciousness
V2,V4,V6: Beauty consciousness
To minimize factor loading

We multiply 2 martrices along with its transpose (matrix will be if PC and Variables)
If any of the variable is less than 0.5 factor loading in all the components or factors.
Then we will6 have to check the variable if we have to remove or not.
RC1 RC2 RC3
V1 .4 .3 .1
V2 .5 .5 .2
V3 .4 .5 .1
To remove the variable we first check the communality .If communality value is less than 0.5 ,go to
the initial correlation matrix , remove that variable and re-run the
Communality of each variable should be atleast 0.5
Better result: Cross loading is minimized. Loading is 0.5 minimum
Each variable should have a communality of 0.5

The minimum cumulative variance should be minimum 60%
1. When we need to find important factors that affect consumer decision

3. Explain the factors based on the variable clubbed in the segments
2. We need to do behavioral analysis
We give answer in segments (Answer is: these are two segments BC and HC)
Based on the two segments explain 4P’s of marketing.

After the segmentation is done
Read the method and hypothesis part of the journal

Define the population (all internet population) what is the popn
determine the sampling frame (online communities). LIST OF MBA STUDENTS UNIVERSITY
Sampling unit ( Post graduate managemtn students), sincerity
Samole size: Ratio of sample size should be 20:1 world class research
No.of items, questions, variables: sample =1:20
If 6 questions are there 120 samples should be there
variables means,
face validity of 40 variables ask if they are relevant
Research design: popn, sample frame, sample unit, sample size,

from 40 we reduce to 30
now the sample size should 30x20= 600
But we can never keep the sample size less than 200
Do two different homogenous group
50 homogenous groups are there
Simple Random Sampling: PROBABBILSTIC IN NATURE

Each and every in the popn has equal chance of being selected in the sample
Simple Random Sample is never possible. Since we do not have access to entire popn , so SRS is
never possible
sampling techniques: Convenience and judgement sampling (NON PROBABILITY SAMPLING

TECHNIQUES)
When we get the sample after convenience and judgemental we use simple random samplimg to
randomize the collected sample
Convenience sampling + judgement sampling (time place)

sampling techniques
define popn
det samling frame
select sampling tech
det sample size
execite the sampling process
Put 4P’s decision
Secondatr source
analysis
stiory

Session 3,4

Uploaded by

Copyright:

Available Formats

Session 3,4

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

Session 3,4

Uploaded by

Copyright:

Available Formats

To identify min no.

segments that explain max variance in data

Bartlett test the hypothesis if there is any correlation

H2: the amount of variance being explained by total no. of factors

Which variable will fall in which component?

The threshold value of factor loadings should be 0.5

We should minimize the cross loading

Here improvement in results means minimization of cross loading

To minimize factor loading

Better result: Cross loading is minimized. Loading is 0.5 minimum

Each variable should have a communality of 0.5

1. When we need to find important factors that affect consumer decision

Based on the two segments explain 4P’s of marketing.

Read the method and hypothesis part of the journal

If 6 questions are there 120 samples should be there

face validity of 40 variables ask if they are relevant

Research design: popn, sample frame, sample unit, sample size,

Simple Random Sampling: PROBABBILSTIC IN NATURE

sampling techniques: Convenience and judgement sampling (NON PROBABILITY SAMPLING

Convenience sampling + judgement sampling (time place)

Put 4P’s decision

You might also like