Through some persistence I figured out to use a "Where" line within the Proc allows splitting.  Here's the details of the code.  And below this I have another Proc PLS for      where yr='2017';   /* Running Model By Year to separate out 2014 from 2017 */ /* Global Model with all bands at R1 */ proc pls data=splitplsr2 cv=random(seed=12345) cvtest varss plots=(diagnostics dmod scores ParmProfiles VIP XLoadingProfiles); by yr; model meas = R1Avg_412_d.......etc. R1Avg_917_d / solution; output OUT=test1; ods output VariableImportancePlot=vip; title 'Global PLS Full Model'; run; /* Model with Factors chosen */ proc pls data=splitplsr2 nfac=3 varss plots=(diagnostics dmod scores ParmProfiles VIP XLoadingProfiles); by year; where yr='2014'; model meas = R1Avg_412_d....etc. R1Avg_917_d / solution; output OUT=outfile predicted=predR1f press=pressR1f yresidual=yresidR1f ; ods output VariableImportancePlot=vip; title 'Global PLS Full Final Model'; run;   ... View more

@Daisy2 wrote: Thanks for the quick reply. After thinking about your question, I think it's most appropriate to do the former and not the latter - not use the records at all. Where do I put your suggested code? Before Proc SurveySelect? (I'm pretty new to SAS) such as Where not missing(nce); proc surveyselect data=....?? or within the procedure? In the procedure:   Proc surveyselect data=   <other proc options>;    where not missing(nce);   <remaing surveyselect statements>   Actual order is not critical but by placing it near the start of the procedure it is easier, IMHO, to remind me I was filtering data for some reason when I come back to the code later. In fact I would likely insert a comment such as /* Do not want the missing values in this set to be considered as a Cluster for selection*/     ... View more

data want; set have; NCMeas = NC; if Selected=0 then call missing(NCMeas) ; run; alternative code: data want; set have; if Selected=0 then call missing(NCMeas) ; else NCMeas = NC; run;   ... View more

PROC PLS and PROC PLM do not work together.   You could split the data, do Cross-Validation on the training data set to determine the number of dimensions, then use the validation data to see how it compares to the training data. That's something you'd have to program yourself, and honestly, I have never done that with PROC PLS. I have done this with a macro I wrote that does Logistic PLS.   Some background: Partial Least Squares was developed way outside mainstream statistical academia, and that's why there is little in the way of hypothesis testing and little in the way of confidence intervals and no SEP or similar. I'm sure there are people who have tried to add this capability Partial Least Squares, but it hasn't made its way into SAS. Nevertheless, the success of Partial Least Squares, with thousands of published papers, handling situations of high multicollinearity of the X variables that are really difficult to handle via other methods is impressive, and somewhat indicates that the statistical hypothesis testing/confidence interval type of analyses are not really necessary. But that makes PLS somewhat uncomfortable for people who are used to hypothesis testing and confidence intervals, and so you need to be aware of this. ... View more

Since the base model is nested within model B, you can get a likelihood ratio test for the significance of adding the B parameter by subtracting the -2 log likelihood for the base model from the same factor of model B.  This should give the same probability as the F test for B in the B model.   But somehow, I don't think this is quite what you are looking for.  I suspect you want to know something about the adequacy of the full models.  This is where information criteria are helpful.  In this case, as there are a fairly substantial number of parameters, I would suggest looking at the corrected AIC values for the two models, and selecting the one with the smaller value.  See  Burnham, K. P.; Anderson, D. R. (2002), Model Selection and Multimodel Inference: A practical information-theoretic approach (2nd ed.), Springer-Verlag ., or for a quick overview, look at the Wikipedia article here.    SteveDenham ... View more