Output Graphics and Action Buttons | Pattern Discovery | MDS Results

MDS Results
The MDS Results tab is shown below:
The MDS Results tab contains the following elements:
Plot of Badness-of-Fit Criterion versus Number of Dimensions
This Overlay Plot should slope downward as the number of dimensions increases. Use it to determine an adequate number of dimensions by looking for the place where the plot begins to flatten, as you might use a Scree Plot for principal components analysis.
Plot of Distance Correlation versus Number of Dimensions
This Overlay Plot slopes upward as the number of dimensions increases. You can use this plot in the same way by finding the point at which the plot begins to flatten.
The following drill-down buttons are available on this tab. Select points in either plot and click the desired button.
Plot Residuals : Highlight a point on one of the plots and click Plot Residuals to plot the residuals from the MDS model fit. Study these to make sure the fit is adequate for a specific number of dimensions. The residuals should ideally look like randomly scattered points around zero. If not, try using a higher number of dimensions.
Display Clustered Heat Map, 2D, and 3D Scores : Highlight a point on one of the plots and click Display Clustered Heat Map, 2D, and 3D Scores to plot the computed MDS coordinate scores using a multivariate 2D scatterplot matrix, a 3D spinning plot (for three or more dimensions), and a clustered heat map of the scores.
Note : Clicking a point on one plot automatically selects the point on the other plot.