#
Mathematical Science in Data Science

Contents of Main
are used my classification of the method.
But this is not clear.

Relationship between the method and the purpose is not "one to one" or "one to many."
It is "many to many."
Because of that relationship is not easy to understand.
The image of the system of methods is the figure in this page.
Processes of the methods are key points to understand.

##
Processes of the Methods

The common points of mathematical models are "Cleansing", "Model Building" and "Model Using."
This classification is useful to understand the relationship among methods and purposes.

###
Cleansing

The first step is raw-data observation.

By the observation, we find
outlier, missing value
,
significant figures
and so on.

The second is to make
Graphs of Disribution
and
line graphs.
We understand the data deeply.

###
Model Building

For example, in
multi-regression analysis
, "Model Building" is
selection of variance
using
linear formulation
.
We look for the sets of X variables which make the most suitable formulation.

###
Model Using

By "Model Building", the formulation clearly expresses the data set is made.
We use the formulation.

##
Relationship between Purposes and Methods

If the purpose is cause-and-effect analysis,
**information that we find on the process of making formulation**
is important.

If the purpose is prediction validity of the formulation is important.
But it is not important for the use of cause-and-effect analysis.

NEXT Softwares of Data Science