TomTom Tracker

Re: How to modify a stored fitted model before using if for scoring

TomTom — Tue, 26 Nov 2013 00:55:03 GMT

Hi Steve

Which values to use for the calculations of x_mean was also giving me a bit of a headache. Thanks for your thoughts, very useful.

Thanks also for the datastep solution to my problem. It's not that easy, I think, that's why I wondered whether I can fiddle with a model in the item-store. Anyway, it looks like nobody got an answer to my question (or the experts are too busy) so I'll have a try at your datastep solution.

Re: simulatiom data with sas/iml

TomTom — Mon, 25 Nov 2013 08:54:56 GMT

Has eps been set to a value?

It might be helpful to look at the output in the log file directly following the line starting with ERROR.

Re: How to modify a stored fitted model before using if for scoring

TomTom — Sat, 23 Nov 2013 08:31:20 GMT

Well, what intend to do is actually more like adding 1 to each variable, however, I thought it would be neat to do this via the stored model and this got me generally interested in whether I can modify these item-store objects.

I essentially want to do what Joyce et al did in in this paper http://www.ncbi.nlm.nih.gov/pubmed/12407470

They calculate adjusted mortality rates.Their correction is done by subtracting the observed value of a confounder from the mean value of this confounder in the dataset and adding this term to the observed outcome (mortality rate). So for one independent variable the adjusted rate is given by

y_adj = y_obs + b (x_mean - x_obs)

How to modify a stored fitted model before using if for scoring

TomTom — Thu, 21 Nov 2013 16:25:08 GMT

Hello,

I have the following problem:

I fitted a model using proc genmod. I stored this model in the item-store using the STORE statement and I now intend to use this model to score a different dataset using proc plm. However, I would like to modify the stored model before using it for scoring. Is it possible to do this? My modification is something like (a bit more complicated actually): add 1 to each parameter. Naively I tried to access the stored model in a data step which doesn't seem to be possible.

Many thanks for your help

Generating numbers with given stepsize and range

TomTom — Thu, 26 Jan 2012 16:15:57 GMT

PROC IML;

i = DO(2,45,4);

PRINT i;

QUIT;

All of the questions you are asking on the SAS/IML forum are rather basic. May I suggest that you buy (or borrow) Rick's book to properly learn IML from scratch?

The simple statistical functions in SAS/IML

TomTom — Thu, 19 Jan 2012 00:12:52 GMT

The problem is that the STD function has only been added to IML in SAS v9.3, you are using an earlier version. Therefore you are calling the SAS/base function STD which has different arguments. Here is its documentation

http://support.sas.com/documentation/cdl/en/lefunctionsref/63354/HTML/default/viewer.htm#p1fani7wj5h6ihn1j1kijj17ylpq.htm

Same problem with the VAR function although this function has been added at some point earlier. I am using SAS v9.23 and for me VAR is available in IML and STD isn't.

Re: Multivariate Normal Tail probabilities

TomTom — Tue, 10 Jan 2012 21:01:52 GMT

Thanks for your reply. I was aware of the Genz & Bretz modules but hoped that there was a (hidden) build-in function that I missed.

Multivariate Normal Tail probabilities

TomTom — Tue, 10 Jan 2012 16:20:02 GMT

Hello,

I was wondering whether there is a SAS/IML function for obtaining tail probabilities of a multivariate Normal RV? Rick has blogged extensively about sampling from a MVN (thanks Rick!) but I can’t find anything about tail probabilities.In essence I am looking for what this package provides in R:

http://cran.r-project.org/web/packages/mnormt/mnormt.pdf

Many thanks

exponential function

TomTom — Sat, 29 Oct 2011 14:35:49 GMT

telescopic wrote:

By the way, in IML, can we assign a value as scalar instead to a (1x1) matrix?

1x1 matrices are scalars

http://support.sas.com/documentation/cdl/en/imlug/64248/HTML/default/viewer.htm#imlug_languagechap_sect001.htm

The main problem with your programme above is that in your likelihood module the result of the calculation of EXP(194481) is too big for the computer to handle. I think you also want to pass rows of p as arguments to LogLik and not the whole matrix. I suppose you have seen Rick's recent blog post. If not, have a look at what he did:

http://blogs.sas.com/content/iml/2011/10/12/maximum-likelihood-estimation-in-sasiml/

Re: windows 7 pipe problem

TomTom — Tue, 25 Oct 2011 11:20:59 GMT

Thanks for pointing me to the usage note. Pinning SAS to the task bar resolves the issue.

Re: windows 7 pipe problem

TomTom — Tue, 25 Oct 2011 10:45:23 GMT

Many thanks for your reply. Unfortunately your code doesn’t work. First, %put %sysget(hostname); gives a warning on XP as well as on Win7:

WARNING: The argument to macro function %SYSGET is not defined as a system variable.

The datastep afterwards works on XP but not on Win 7. The error message is similar to the one in my original post:

NOTE:The infile ENVS is:

Unnamed Pipe Access Device,

PROCESS=set,RECFM=V,LRECL=10000

Stderr output:

The handle is invalid.

(More handle invalid lines)
The handle is invalid.

NOTE: 0 records were read from the infile ENVS.

NOTE: The data set WORK.EVARS has 0 observations and 3 variables.

NOTE: DATA statement used (Total process time):

real time 0.14 seconds

user cpu time 0.01 seconds

system cpu time 0.03 seconds

Maybe I should mention that, unlike XP, the Win 7 installation is 64 bit.

windows 7 pipe problem

TomTom — Tue, 25 Oct 2011 10:34:34 GMT

Because I wasn't aware of this macro variable :smileysilly: Using syshostname works, thanks! I still would like to find out what is wrong with my original code, though. Maybe my response to Peter.C gives a lead...

windows 7 pipe problem

TomTom — Tue, 25 Oct 2011 10:30:57 GMT

It gives me the hostname, assuming that what you call 'command prompt' is what I called 'windows console' (cmd.exe).

windows 7 pipe problem

TomTom — Mon, 24 Oct 2011 20:24:37 GMT

Hello,

I have a problem with one of my macros using SAS 9.22 on Windows 7. I amrunning my programmes on different computers in different directories. ThereforeI wrote a little macro which checks the computer’s hostname and according tothis name assigns my library to a certain directory. I do the following to putthe hostname into a macro variable:

FILENAME h PIPE 'hostname';

DATA_NULL_;

INFILE h;

INPUT;

PUT _infile_ ;

CALL SYMPUT('current_com',TRIM(LEFT(_infile_)));

RUN;

This works with Win XP and SAS 9.22. Here is what thelog says:

NOTE:The infile H is:

Unnamed Pipe Access Device,

PROCESS=hostname,RECFM=V,LRECL=256

myhostname

NOTE: 1record was read from the infile H.

The minimum record length was 10.

The maximum record length was 10.

NOTE:DATA statement used (Total process time):

real time 0.09 seconds

user cpu time 0.00 seconds
system cpu time 0.00 seconds

However, on Windows 7 I obtain:

NOTE:The infile H is:

Unnamed Pipe Access Device,

PROCESS=hostname,RECFM=V,LRECL=256

NOTE: 0records were read from the infile H.

NOTE:DATA statement used (Total process time):

real time 0.31 seconds
cpu time 0.04 seconds

So apparently the hostname couldn’t be read. If I type ‘hostname’in the windows console I get the hostname as expected.

P.S. I apologise if this isn't the correct forum section. It seems most appropriate

Re: Problem with time() function

TomTom — Mon, 03 Oct 2011 20:12:00 GMT

Hmm, I missed the second part of Rick's answer. Sorry about that. Actually, I was replying to the notification email I received for his post and the reply in there looks a bit different (i.e. doesn't contain the datetime bit). I'll try the datetime() function - thanks!

To answer the question why the computation takes so long: I am running what I would call a fairly complex Monte-Carlo simulation. By the way, thanks for implementing the submit statement allowing me to call procedures from within IML. I am using it for these simulations and it makes my life so much easier.

Problem with time() function

TomTom — Mon, 03 Oct 2011 12:40:46 GMT

Hi Rick, thanks for your swift reply.

Are you saying that i should put

t0 = time(); /* overwrite t0; we don't need the original value */

immediately before the step 2 computation?

To be honest, I fail to see the difference between your t2 and mine.

In my code t2 = (time() - t0) is an elapsed time, like in yours, but it is the time from start to finish of the whole programme. From this I am subtracting the elapsed time for the first step. Can I not use elapsed times in calculations? I'd greatly appreciate it if you could elaborate.

Problem with time() function

TomTom — Mon, 03 Oct 2011 12:07:37 GMT

Hello,

I have a problem with the time() function. My code looks like this:

PROC IML;

t0 = time();

/* Step 1 computation */

t1 = time() - t0;

/* Step 2 computation */

t2 = (time() - t0)-t1;

msg = "Step 1: _x seconds Step 2: _y seconds";

sX = putn(t1, "BEST5.");

sY = putn(t2, "BESTD9.");

msg = tranwrd(msg, "_x", sX);

msg = tranwrd(msg, "_y", sY);

PRINT msg[label="Time evaluation"];

QUIT;

The time for step 1 is shown correctly but step 2 usually is not. I get all sorts of numbers for t2 including negative ones. I suppose the problem is that step 2 takes considerably longer than step 1 (20 to 30 hours for step 2, <1 hour for step 1). Is the time function unable to handle such long timeframes or could it be that I just don’t get the format in the putn function right? I tried several BESTxx.x values but none seemed to work. The code works fine for 2 quick computations.

Many thanks