BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
Solph
Pyrite | Level 9

I've a data set of 70 cases, and I have an equation involving 17 variables from the data that I'd like to test on. Specifically I'd like to test if

1) changing +/-5% magnitude of each variable ^one at a time^, what would the variation for the Y value on the left of the equation and changing which variable would produce the largest variation.

2) changing +/-5% magnitude of each variable ^one at a time^, summing up the Y value and calculating the proportion of each case (so the sum if 100%), what is the variation for share % and changing which variable would produce the largest variation in the share?

3) The same as #2, except changing +/-5% magnitude of each variable ^one at a time^ and one case a time, which variable would produce the the largest variation for the overall data and which case is most impacted by this change. (Hope this question makes sense).

 

I got some great tips from and at the Programming forum  https://communities.sas.com/t5/Base-SAS-Programming/Code-to-perform-simulation/m-p/279203#U279203 and were suggested to post here. Ksharp gave me a code for #1 but it has a bit of error. How do I fix?

And how do I approach #2?

 

Doing simulation in SAS is all new to me and it's a bit urgent for work. (I'll definetely need to read more on the subject for future work).

 

data have; input
            ID A B C D E N1 N2 N3 N4 N5 N6 D1 D2 D3 D4 D5 D6;
datalines;
1  100 10 0.5 15 10  35 8 6 4 2 1 39 8 5 5 1 2
2  200 11 0.3 25 15  35 8 6 4 2 2 42 7 7 4 2 3
3  300 20 0.4 15 14  40 8 6 4 2 3 53 9 6 5 3 2
4  150 19 0.1  5 19  27 8 6 4 2 4 35 4 9 7 4 4
5  400 40 0.5 35 20 50 8 6 4 2 6 79 17 4 9 1 5
6  900 50 0.4 55 15  70 8 6 4 2 9 85 5 9 4 2 11
7  600 30 -0.5 20 29  60 8 6 4 2 2 62 16 7 5 4 10
8  500 30 0.2 25 17  50 8 6 4 2 3 60 4 9 8 2 5
9  400 30 0.4 19 16  44 8 6 4 2 1 49 8 4 4 5 9
10  700 45 0.3 39 14  34 8 6 4 2 1 38 3 2 5 4 12
;
run;

proc iml;
use have(keep= A B C D E);
read all var _num_ into have[c=vnames];
close;

use have(keep= N1-N6);
read all var _num_ into N;
close;

use have(keep= D1-D6);
read all var _num_ into D;
close;

Parm_N={1 -1 -1 -1 -1 -1};
Parm_D={1 -1 -1 -1 -1 -1};
magnitude={-1.05 1.05};
Y=j(nrow(have),1,.);
id=j(ncol(magnitude)#ncol(have),1,blankstr(40));
mean_std=j(ncol(magnitude)#ncol(have),2,.);
mattrib mean_std c={mean std};


idx=0;
do i=1 to ncol(magnitude);
 do j=1 to ncol(vnames);
  Parm_have=j(1,ncol(have),1);
  Parm_have[j]=magnitude[i];  
  do k=1 to nrow(have);
   temp=have[k,];
    wgt=(N[k,]*Parm_N`)/(D[k,]*Parm_D`);
   wgt=choose(wgt>1,1,wgt);
   Y[k]=(Parm_have[1]#temp[1]#
         Parm_have[2]#temp[2]#
         (Parm_have[3]#temp[3]+1)
         -Parm_have[4]#temp[4]
         -Parm_have[5]#temp[5])#wgt; 
  end;    
  idx=idx+1;
  id[idx]=char(magnitude[i])+'*'+vnames[j];
  mean_std[idx,1]=mean(Y);
  mean_std[idx,2]=std(Y);
 end;
end;

print id mean_std[l=''];
quit;

 

Thanks in advance.

1 ACCEPTED SOLUTION

Accepted Solutions
dougc
SAS Employee

If you are looking to vary the magnitude of each variable by +/- 5%, then you need to change

 

magnitude = {-1.05 1.05};

 

to

 

magnitude = {0.95 1.05};

 

 

If I understand correctly what you are trying to do with 2), you should be able to add

 

Y = Y/sum(Y);

immediately following your loop over k.  Leaving everything else the same *should* give you the answer for 2).  If I have misunderstood what you are trying to accomplish, please elaborate further.

View solution in original post

2 REPLIES 2
dougc
SAS Employee

If you are looking to vary the magnitude of each variable by +/- 5%, then you need to change

 

magnitude = {-1.05 1.05};

 

to

 

magnitude = {0.95 1.05};

 

 

If I understand correctly what you are trying to do with 2), you should be able to add

 

Y = Y/sum(Y);

immediately following your loop over k.  Leaving everything else the same *should* give you the answer for 2).  If I have misunderstood what you are trying to accomplish, please elaborate further.

Solph
Pyrite | Level 9

Thanks a lot. Will fix and apply. 

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

Multiple Linear Regression in SAS

Learn how to run multiple linear regression models with and without interactions, presented by SAS user Alex Chaplin.

Find more tutorials on the SAS Users YouTube channel.

From The DO Loop
Want more? Visit our blog for more articles like these.
Discussion stats
  • 2 replies
  • 811 views
  • 0 likes
  • 2 in conversation