Significant model? I have 1,000 records to work with and the model is misclassifying 35%, so maybe not. Not sure I get the comment about adding more noise. Even if I had mostly noise, why would adding more variables result in a worse decision tree? ... View more

I had a similar problem and I was made after several attempts. The problem most like is that one or more variables are not yet in the data file. Take time to cross-check that. Thank you. ... View more

Use the CODE statement in either proc and examine the output.    The initial thread I linked too walked through the solution for LOGISTIC regression in detail so I'm not sure what I could add besides replicating it specifically for your exact values.  ... View more

If your ultimate objective is to project out, then I would really urge you to look into the SAS/ETS procedures--ARIMA and MODEL may be appropriate.  Extrapolating beyond the range of existing data is extremely dangerous for ordinary least squares or maximum likelihood methods, such as those used by SAS/STAT procedures.  However, it can be done.  Append records to the end of your data, indicating months and with growth_rate set to missing.  For the two time points you mention, it would mean adding 120 . 240 . to your existing dataset.  Now whatever procedure you use will fit based on complete data, but OUTPUT datasets can be set up to contain predicted values and confidence intervals. I would do a search on this site for "hockey stick" and "PROC NLIN" as a way to fit this data. Steve Denham ... View more

Hi: You have 2 choices: 1) Use Univariate internal OUTPUT syntax to name the variables listed in your VAR statement or 2) Use ODS OUTPUT syntax to get the "BASICMEASURES" output object that will contain all the info for the variables listed in the VAR statement. cynthia [pre] ** 1) Use Univariate OUTPUT statement to create multiple variables; ods trace on /label; proc univariate data = sashelp.class; var age weight height; output out = my_output_dataset median =medage medwt medht mean =mnage mnwt mnht mode =modage modwt modht ; run; ods trace off; ods listing; proc print data=my_output_dataset; title '1) Using Internal OUTPUT statement'; run; ** 2) use ODS OUTPUT statement; ods output basicmeasures=work.all_basic; proc univariate data = sashelp.class; var age weight height; run; proc print data=work.all_basic; title '2) Using ODS OUTPUT'; run; [/pre] ... View more

In the classic regression sense, you can do this with interaction terms in one model. As a concrete example, say that X=baby diapers and Y=Viagra. So, you'd expect X-people to be mostly younger and female and the Y-people to be mostly older and male. If X and Y are 0/1 coded, a model might look something like $$spent = age sex x y age*x age*y sex*x sex*y Doc Muhlbaier Duke ... View more

recode the variables so they are in the order you want. ... View more

Ah, yes. That looks familiar. Worked like a charm. Thank you! ... View more

good ... View more

Yes and no. It is a different procedure and generates a single test for the hypothesis that there is no effect on the combination of the two dependent variables. In the multivariate context, GLM also allows one to test multivariate linear combinations of the dependent variables, providing a different approach to repeated measures analyses than you would get in MIXED modeling. As a side effect, it also does the two ANOVAs, with a caveat about missing values. GLM does listwise deletion, so you will get different ANOVA results if the dependent variables have different missing value patterns than you would get by running the GLM procedure once for each dependent variable. Doc Muhlbaier ... View more

PROC FREQ will do both of these. 1) is the natural way. PRC FREQ; TABLES team*cancelled_flag/chisq; RUN; 2) requires data in slightly different format team cancelled_flag count a y 40 a n 60 b y 50 b n 70 Then the proc looks like PROC FREQ; TABLES team*cancelled_flag/CHISQ; WEIGHT count; RUN; ... View more

[pre] proc sort data=indata by person_id date; data expanded; set indata; cmonth = month(date); run; quit; data expanded; set expanded; by person_id cmonth; retain month sum; if first.person_id then month = 0; if first.cmonth then do; sum = 0; month+1; end; diff = cmonth - lag(cmonth); if diff < 2 then sum+amount; else do i=1 to diff; amount=0; output; month+1; end; if last.cmonth then do; amount = sum; output; end; drop cmonth sum diff i; end; [/pre] ... View more