Partial Least Squares Plots with PROC PLS

This example is taken from the section Getting Started: PLS Procedure of Chapter 70: The PLS Procedure. The following statements create a SAS data set that contains measurements of biological activity in the Baltic Sea:

data Sample;
   input obsnam $ v1-v27 ls ha dt @@;
   datalines;
EM1   2766 2610 3306 3630 3600 3438 3213 3051 2907 2844 2796
      2787 2760 2754 2670 2520 2310 2100 1917 1755 1602 1467
      1353 1260 1167 1101 1017         3.0110  0.0000   0.00
EM2   1492 1419 1369 1158  958  887  905  929  920  887  800

   ... more lines ...   

;

The following statements run PROC PLS:

ods graphics on;

proc pls data=sample cv=split cvtest(seed=104);
   model ls ha dt = v1-v27;
run;

By default, the procedure produces a plot for the cross validation analysis and a correlation loading plot (see Figure 21.8).

Figure 21.8: Partial Least Squares

Partial Least Squares
External File:images/plsg1a1.png