About neilxu

neilxu · ‎06-28-2017

Hi All, I used SAS/OR proc optmodel before in order to find a solution. But that's for a single number. Now I need to randomly pick 100 records from 1000 records. There is another constaint, the sum of these 100 records must be a ceterain number (K). So the constaints are: sum of the total number of records is 100 and the sum of records is K. Where can I start? I look up some of examples but I cannot find anything similar. Thanks

neilxu · ‎06-20-2014

I am using PROC SGPLOT; BY segment; series ...... However the title is truncated if the value in segment variable is very long. Anyway to solve this problem? I wrote a macro to take care of this, but everything I have to plot each segment which is not very efficient. Thanks

neilxu · ‎06-12-2014

Thank you so much.

neilxu · ‎06-10-2014

If, for example, I have monthly sales number for different department (total 10 departments). Now I want to look at the trend of these sales based on departments. However 10 is too much so naturally clustering/grouping comes to my mind. I saw the example of proc similarity to cluster time series and followed it to create the clusters. Now my question is I can use proc corr to group these departments, right? what is the benefit to do proc similarity? Thanks

neilxu · ‎06-04-2014

if I specify selection=lasso(stop=ADJRSQ); then SAS stop in 2 steps and show: Selection stopped at a local maximum of the AdjRSq criterion. If I specify selection=lasso(stop=SBC);then SAS stop in 2 steps and show: Selection stopped at a local minimum of the SBC criterion. I only get 2 variables. The AdjRSq is pretty low in either test unless I specify steps=20. With STEPS option, the AdjRSq increases, However the purpose of using lasso is to avoid overfitting. I look at the variables and I believe STEPS is giving me overfitting result. Look at the correlation between those variables, don't believe all of them are strongly correlated. Thanks for your help

neilxu · ‎06-03-2014

This is my first time to use glmselect with lasso options. However the procedure ends very quickly, always 2 steps. I changed the STOP options but no luck. And the result is really bad, R^2 is below 0.3. Don't understand why it just stops. I have more than 200 IV and only 1 DV (50 records). Thanks for you input.

neilxu · ‎04-01-2014

Hi Rob, We tried and it didn't work. We even contact SAS support and raised a ticket. Nor working at this moment. for example, we are minimize the function min z=sum{i,j} ( x[i,j] - y[i,j] ) **2 the constraints are, for example: con condition1{i}: sum{j} x[i,j]=1 now we want to introduce expmatrix(x,3,z) and make sure z is satisfied with some other conditions. Can you give us an example how to use SUBMIT block? Thanks

neilxu · ‎03-31-2014

Hi Rob, thanks for your help. Unfortunately, we are still using SAS9.3. My thought is: keep the original OPTMODEL, add one more constraint in SUBMIT block. Can you show me how to use SUBMIT instead of FCMP, for example? Thanks

neilxu · ‎03-28-2014

Thanks Rob. The problem is in current OPTMODEL, one of our SAS programmers cannot call expmatrix function. So I rewrote everything in IML with Rick's help. However I am wondering if possible to call expmatrix function in SUBMIT/ENDSUBMIT block. My guess is it's highly possible. So you are right, the youtube example uses SUBMIT to generate output info that can be used as a constraint. My specific question is: that example is too complicated. Do you have some simple examples I can start with? Thanks

neilxu · ‎03-28-2014

Thank you Rob. I did read the part, however the SUBMIT/ENDSUBMIT block is not a constraint. That block takes the outputs and plots a graph. What I am looking at is to embed a constraint in SUBMIT/ENDSUBMIT block. Any simple example? Thanks

neilxu · ‎03-28-2014

Thank you Rick. by the way, it seems IML is faster than OPTMODEL.

neilxu · ‎03-27-2014

Saw a youtube video yesterday. Ed Hughes showed how to use SUBMIT block with PROC CLP to add one more constraint in OPTMODEL procedure. His example is too complicated and I have hard time to understand all details. Is there any simple case which I can start with? User guide and course note don't have much info. Thanks

neilxu · ‎03-27-2014

Hi Rick, the following is the sample codes I have. (I only keep one penalty value) proc iml; start penalty1(m,a); v=m-a; return(choose(v>0,0,v##2)); finish; start sumsqr(a) global(m1); d1=a[1:nrow(a)-1,1:ncol(a)]; p1=ssq(penalty1(d1,0.01)); ccc=(a-m1)##2; zzz=sum(ccc)+p1; return (zzz); finish; aaa={0.8 0.15 0.04 0.01, 0.1 0.8 0.09 0.01, 0.05 0.1 0.8 0.05, 0 0 0 1}; m1={0.7 0.25 0.04 0.01, 0.2 0.7 0.09 0.01, 0.05 0.2 0.7 0.05, 0 0 0 1}; test=sumsqr(aaa); print test; x0=j(4,4,0.1); opt={1,4}; call nlpnra(rc, result, "sumsqr", x0, opt); quit;

neilxu · ‎03-27-2014

tried test=OBJECTIVE_FUNCTION(x0); no error. tried call nlpnra(rc,result,'fun',AAA,opt); no luck, same error message. however, if I change x0 to x0=j(1,4,0.1); well, I got this ERROR: (execution) Invalid subscript or subscript out of range. once again, thanks for your help

neilxu · ‎03-27-2014

I tested my codes as followings: START PENALTY FUNCTION; FINISH; START OBJECTIVE_FUNCTION(AAA); NEW OBJECTIVE FUNCTION=ORIGINAL FUNCTION + PENALTY RETURN(ZZZ); FINISH; AAA={ ............}; test=OBJECTIVE_FUNCTION(AAA); print test; and the result is what I expect. So this part is correct. now I remove AAA and test part, then write: x0=j(4,4,0.1) opt={1,4}; call nlpnra(rc,result,'fun',x0,opt); I got the error saying "NLPNRA call: Error in argument X0". Either I should not use NLPNRA function nor the syntax of X0 is wrong. Which one? Thanks

Online Status	Offline
Date Last Visited	‎07-05-2017 09:50 AM

find a vector solution from two constraints

Title is truncated in SGPLOT with BY

Re: Proc Similarity to cluster time series data

Proc Similarity to cluster time series data

Re: glmselect with lasso options ends only 2 steps

glmselect with lasso options ends only 2 steps

Re: How to use submit block to control constraints

Re: How to use submit block to control constraints

Re: How to use submit block to control constraints

Re: How to use submit block to control constraints

Re: find a vector solution from two constraints

find a vector solution from two constraints

Title is truncated in SGPLOT with BY

Re: Proc Similarity to cluster time series data

Proc Similarity to cluster time series data

Re: glmselect with lasso options ends only 2 steps

glmselect with lasso options ends only 2 steps

Re: How to use submit block to control constraints

Re: How to use submit block to control constraints

Re: How to use submit block to control constraints

Re: How to use submit block to control constraints

Re: converting SAS OPTMODEL to SAS/IML

How to use submit block to control constraints

Re: converting SAS OPTMODEL to SAS/IML

Re: converting SAS OPTMODEL to SAS/IML

Re: converting SAS OPTMODEL to SAS/IML