I don't think this is possible for PROC HPFOREST (which creates a forest of decision trees). However, for a single tree model, you can get this information from PROC HPSPLIT.  ... View more

Thank you for posting the log. I'm now able to reproduce your results. My machine defaults to using 4 threads for the computation, but I needed to use 2 threads to match your output. Note that you do so using the performance statement as follows: proc hpsplit data=small seed=5 CVCC PLOTS=CVCC cvmethod=random(2) CVMODELFIT NODES=DETAIL; class y; model y = x; GROW entropy; PRUNE costcomplexity; performance nthreads = 2; run; I suggest that you raise this issue with SAS technical support (make sure they also use nthreads = 2). I think what is happening is the selected tuning parameter is the one with the smallest average squared error based on cross validation. In the documentation for the 'CVMETHOD= random' statement, it says "The average ASE across the k trees is the cross validation error for that set of trees .... the parameter that has the minimum cross validated error is used as the best parameter value."   However, to your point, you are fitting a classification tree, and the error metric for a classification tree is commonly the misclassification rate.  In fact, the documentation for prune statement, says "The error metric is misclassification rate for classification trees". I think it is worth confirming with technical support "that PROC HPSPLIT is working as intended.  ... View more

Can you post the log for this program? I'm not reproducing this behavior, but I'm not using SAS on Demand for Academics.  ... View more

I've moved this post to the IML board, which more closely aligns with this question. As pointed out by a previous answer, from the IML documentation, the eigenvalues for a non-symmetric matrix are sorted first by their real parts, then by the magnitude of their imaginary parts. You can get sorted eigenvalues from numpy using the numpy.linalg.eigh() function, but note that that function sorts eigenvalues in ascending order.  ... View more

Agree with the post from @StatDave . I will add on that one of the reasons people sometimes use an identity link to model binary response data is that you are directly modeling the probability of the event. That is, the model coefficients can be interpreted as estimates of the difference in the probability of the event. Whereas, for the logit link, you are modeling the log odds of the event. The model coefficients are therefore estimates of log odds ratios, and some people find log odds ratios hard to interpret.  ... View more

This is a nice paper that helps explain some of the different options you have regarding random vs. repeated statements :  Advanced Techniques for Fitting Mixed Models Using SAS/STAT Software . The syntax you have for your random statement is not quite right. Generally, you specify one or more model effects as random, and then use the 'subject' keyword to define the subjects to which the random effects apply. For instance, you could fit a random intercept for each subject with something like: random int / subject = PT_ID;   ... View more

Answering your questions in order:   1. You can do either, but I think the more common approach in genetic studies is to fit 300 models, each with just one of the SNPs as a predictor variable.   2. Yes. You calculate m using the simpleM method, and then declare 'statistically significant' p-values less than 0.05/m.  ... View more

You should not compute a post-hoc power analysis. A power analysis is only useful prior to the study to estimate a reasonable target enrollment number. Once the study is complete, you can either run the superiority test or not. You will have a better idea if reviewers in your field will accept a superiority test for a trial designed to detect non-inferiority.  ... View more

Take a look at the example here: https://go.documentation.sas.com/doc/en/pgmsascdc/v_048/statug/statug_psmatch_examples04.htm . For binary classification variables, in your case, for the SEXE variable, the computed value is the difference in the proportion of the first ordered level of the variable between treated and control groups.  ... View more

In the code specified, you are including a random effect for each animal in your study. In general, if there are 282 animals with non-missing phenotypic records, you will get 282 random effects. I don't have access to the book you have referenced, so I'm not sure if the authors did some extra step to account for the missing data in the other 24 animals.  ... View more

Can you give more specific details about your study? In general, survival outcomes should be modeled using a procedure that is tailored to survival data, such as PROC PHREG. However, those kinds of procedures require that you have the survival (or censoring) time for each subject studied.  ... View more