Hierarchical clustering
·
Connectivity based clustering, also known as
hierarchical clustering, is based on the idea of objects being more related to
nearby objects than object which are not related
·
If
objects are less than 50 then hierarchical clustering is preferred.
There three important things we
should know before clustering
1) Selection
of variable which will be determined from the objective of group
2) Distance
measurement Criteria
3) Clustering
Criteria
Important rules while analyzing data
Rule I: Look at the data
from different perspective
Rule II: Combine results
from different analysis
The OLAP (Online Analytical
Processing) is one of the tool used for combining the result from different
analysis
·
The
OLAP (Online Analytical Processing) Cubes procedure calculates totals,
means, and other statistics for continuous summary variables within categories
of one or more categorical grouping variables.
·
The
cube allows you to look at different combinations of the categorical variable’s
values to help pinpoint differences in the statistics and focus on specific
values for report purposes.
·
The
OLAP Cube is a good way to focus on data subgroups because a
separate table layer can be created for each possible combination value of the
grouping variable’s values
Example: Total and average sales for
different regions and product lines within regions.
Combining
result from different analysis gives new dimensions to the analysis of data. For
example the data we analyzed cell centre and we found how even revenue/profit
combination was different for calls and SMS. This kind of analysis might have
helped companies to plan their strategy for example Vodafone is only focusing
on Value added services while advertising.
Ashitosh Mohite
14012
No comments:
Post a Comment