How can I graph my results from MVE or MCD analysis?

[MECT]

Hello,

I have just discovered SAS/IML while looking for a means to identify outliers in multivariate data with 3 dimensions. It looks as if either the Minimum Volume Ellipsoid (MVE) or Minimum Covariance Determinant (MCD) may be just what I need. I have looked at the Robust Regression examples within the SAS documentation as well as "Detecting outliers in SAS: Part 3: Multivariate location and scatter" (Wicklin, 2012). Using those sources I have managed to run both MVE and MCD and get estimated outlier, however, I cannot graph the data. I have pasted below the example in the SAS documentation using the stackloss data where a subroutine SCATMVE is used. It seems to be in a library file called robustmc.sas which I do not seem to have. I am running SAS version 9.3 (although I can upgrade to 9.4). I just feel like there are some key points missing in the example and I would be grateful for any assistance.

Thank you

The following specification generates three bivariate plots of the classical and robust tolerance ellipsoids. They are shown in Figure 12.5.5, Figure 12.5.6, and Figure 12.5.7, one plot for each pair of variables.

optn = j(9,1,.); optn[5]= -1;
vnam = { "Rate", "Temperature", "AcidConcent" };
filn = "stlmve";
titl = "Stackloss Data: Use All Subsets";
%include 'robustmc.sas';
call scatmve(2,optn,.9,a,vnam,titl,1,filn);

[Rick_SAS]

I know longer have SAS 9.3 installed, and I don't remember the details from that era, but the current SAS/IML documentation answers your question. In SAS 9.4 you can  %INCLUDE the robustmc.sas file and then use the functions. The example shows the details, but it begins with 

%include sampsrc(robustmc);   /* define graphing modules */
proc iml;
load module=_all_;            /* load graphing modules */

Although I don't remember the 9.3 details, I think you can follow the instructions for "Accessing documentation and sample programs" to get to the 9.3 programs. If necessary, you can copy the programs onto your local computer or cut-and-paste directly into your program.