Tools CorrelationMap: Difference between revisions

From Eigenvector Research Documentation Wiki
Jump to navigation Jump to search
imported>Jeremy
No edit summary
imported>Jeremy
No edit summary
 
(4 intermediate revisions by the same user not shown)
Line 1: Line 1:
__TOC__
[[TableOfContents|Table of Contents]] | [[Tools_ModelRobustness|Previous]] | [[TableOfContents|Next]]
[[TableOfContents|Table of Contents]] | [[Tools_ModelRobustness|Previous]] | [[Index|Next]]


==Correlation Map Tool==
==Correlation Map Tool==
Line 50: Line 49:
The figure below shows a correlation map with the variables in their original order after default preprocessing methods were carried out on a 200 x 30 DataSet for a simple PCA model. The color and intensity of the off-diagonal elements indicates the correlation among the variables. For example, in the figure below, variable 18 correlates well with variables 7, 8, an d 9 but it does not correlate well with variables 4 and 5.
The figure below shows a correlation map with the variables in their original order after default preprocessing methods were carried out on a 200 x 30 DataSet for a simple PCA model. The color and intensity of the off-diagonal elements indicates the correlation among the variables. For example, in the figure below, variable 18 correlates well with variables 7, 8, an d 9 but it does not correlate well with variables 4 and 5.


::''Example of a Correlation map''
:''Example of a Correlation map''


::
::
Line 60: Line 59:
To zoom in on an area for viewing, click the Zoom in icon [[Image:Zoom_In_icon.png|19x21px]]. and then drag your cursor around the region of interest. A box is formed around the area that being reduced for viewing. To zoom out on area for viewing, click the Zoom Out icon [[Image:Zoom_Out_icon.png|20x21px]] and then drag your cursor around the region of interest. A box is formed around the area that being enlarged for viewing. The figure below shows a region that was reduced for viewing. In this figure, the correlation of variable 18 with variables 7 and 8 is much easier to ascertain.
To zoom in on an area for viewing, click the Zoom in icon [[Image:Zoom_In_icon.png|19x21px]]. and then drag your cursor around the region of interest. A box is formed around the area that being reduced for viewing. To zoom out on area for viewing, click the Zoom Out icon [[Image:Zoom_Out_icon.png|20x21px]] and then drag your cursor around the region of interest. A box is formed around the area that being enlarged for viewing. The figure below shows a region that was reduced for viewing. In this figure, the correlation of variable 18 with variables 7 and 8 is much easier to ascertain.


::''Example of zooming in on a Correlation map''
:''Example of zooming in on a Correlation map''


::
::

Latest revision as of 13:50, 1 September 2011

Table of Contents | Previous | Next

Correlation Map Tool

The Correlation Map is a tool that shows the degree of correlation among the variables after you have loaded x block data. You can show the degree of correlation before you preprocess the x block data or after you preprocess the x block data. In addition, you can show the correlation among the variables in one of three ways:

Option

Description

No variable reordering

Variables are plotted in their original order.

Correlation reordering

Adjacent variables will have the highest degree of positive correlation when variables are regrouped by similarity.

Absolute value reordering

Adjacent variables will have the highest degree of positive correlation and/or negative correlation when variables are regrouped by similarity.

To create a correlation map, on the Analysis window main menu, click Tools > Correlation Map, and then select the way in which you want to order the variables.

The figure below shows a correlation map with the variables in their original order after default preprocessing methods were carried out on a 200 x 30 DataSet for a simple PCA model. The color and intensity of the off-diagonal elements indicates the correlation among the variables. For example, in the figure below, variable 18 correlates well with variables 7, 8, an d 9 but it does not correlate well with variables 4 and 5.

Example of a Correlation map
Correlation Map Preprocessing.png

To zoom in on an area for viewing, click the Zoom in icon Zoom In icon.png. and then drag your cursor around the region of interest. A box is formed around the area that being reduced for viewing. To zoom out on area for viewing, click the Zoom Out icon Zoom Out icon.png and then drag your cursor around the region of interest. A box is formed around the area that being enlarged for viewing. The figure below shows a region that was reduced for viewing. In this figure, the correlation of variable 18 with variables 7 and 8 is much easier to ascertain.

Example of zooming in on a Correlation map
Correlation Map Preprocessing ZoomIn.png