The Partial Correlation Diagram report in the Multivariate platform provides a visual representation of the partial correlations. The diagram contains a node for each variable and an edge for each pair of variables. The placement of the nodes in the diagram is determined by the Node Orientation option. There are two options; circular and eigenvector. The Circular orientation spaces the nodes evenly in a circle (or oval, if the diagram is not square). The Eigenvector orientation spaces the nodes based on the eigenvectors. Each eigenvector is multiplied by the square root of the corresponding eigenvalue to create a set of scaled eigenvectors. The coordinates on the diagram are the scaled eigenvectors of the dimensions specified below the diagram.
Each variable node is connected to the other nodes by a line. The color of the line indicates a positive (red) or negative (blue) correlation between the variables. The width of the line indicates the strength of the correlation between the variables. The thicker the line, the stronger the correlation.
Use the following options to adjust the diagram:
Node Orientation
Changes the orientation of the nodes to be Eigenvector or Circular.
Rotate
Use the rotate button to rotate the diagram.
Abs Partial Correlation
Hides lines whose absolute partial correlation value is below the specified number. You can specify the cut off by entering a value in the box next to Abs Partial Correlation or using the slider bar.
Color Separation
Adjusts the line colors.
Width
Adjusts the width of the lines.
Transparency
Adjusts the transparency of the lines.
Reset
Resets all of the display options back to the default values.
Select Dimension
(Available only for the eigenvalue orientation.) Specifies the eigenvector dimensions to use for the node spacing.
The Partial Correlation Diagram red triangle menu contains the following options:
Eigenvalues
Shows or hides a table of the sorted eigenvalues from the decomposition of the partial correlation matrix.
Eigenvectors
Shows or hides a table of the eigenvectors from the decomposition of the partial correlation matrix.
Show Partial Correlation
Shows or hides the partial correlation values on top of the corresponding arcs on the diagram.
Color Theme
Enables you to select a the color theme for the diagram.
Save Coordinates
Saves the scaled eigenvectors (dimensions) to a new data table. The scaled eigenvectors are the eigenvectors multiplied by the square root of the corresponding eigenvalue. These are the coordinates used in the partial correlation diagram.