About kessler

kessler · ‎09-06-2022

Alexey, If you want to programmatically generate your list of restrictions then you can use the SAS macro language. Here's an example that shows how a RESTRICT statement could be generated from values stored in a data set. Marc data d; do t = 2 to 12 by 2; output; end; run; data _null_; length eqs $ 1024; set d; eqname = 'eq'||left(put(_n_,6.)); call symput(eqname, "lhs = " || t); run; %macro restricts; restrict %do i = 1 %to 6; %unquote(&eq&i) %if &i ne 6 %then, %end; ; %mend restricts; proc tmodel; ... your model code ... %restricts; quit;

kessler · ‎08-29-2022

Alexey, If you're new to PROC MODEL it can be difficult to know the best syntax to express your model. I'd be happy to help if you could clarify what data you have and what you are trying to estimate. If your data consists of many contract prices (price) then your FIT statement (ignoring extraneous options) would just be fit price; How you write out the right hand side of your price model depends on how time, discount factors, amounts (remaining and real) are represented in your INUPUT data set (e.g. what are the rows and columns in your INPUT data set). If you can specify these details it should be possible estimate your model. Marc

kessler · ‎05-04-2020

PROC MODEL requires that your model is specified using a mathematical expression including explicit parameters rather than just using a list of variables. To do a two stage least squares estimation for your model you could use something like the following: proc model data=new; am = p_batie*batie + p_ln_ta*ln_ta + p_ROA*ROA + p_leverage*leverage + p_growth*growth + p_age*age + p_CPA*CPA + p_big4*big4 + p_ibsize*ibsize + p_loss*loss + p_loss_lag*loss_lag + p_liquidity*liquidity + p_INVplusAR_TA*INVplusAR_TA + p_btm*btm; fit am / 2sls outv=vdata vardef=n kernel=(bart,0,); instruments _exog_ employ LD; title '2sls results'; quit;

kessler · ‎01-14-2020

If you want all your results in a single data set you'll have to concatenate them yourself. You could do something like this: proc model data=have; y1=a+b*x; y2=c+d*x; y3=e+f*x; fit y1 / gmm kernel=(bart,2,0) vardef=n; ods output parameterestimates=est1; fit y2 / gmm kernel=(bart,2,0) vardef=n; ods output parameterestimates=est2; fit y3 / gmm kernel=(bart,2,0) vardef=n; ods output parameterestimates=est3; quit; data ests; set est1 est2 est3; run;

kessler · ‎08-05-2019

You can restrict your solution conditionally by using the OPTIMIZE option on the SOLVE statement. Below is an example of a simple case where only the solution for the 9th observation is restricted to be less than or equal to 8. Please let me know if you had something else in mind. data d; do t=1 to 10; output; end; run; proc model data=d; y = t**2/10; solve y / optimize out=o; restrict ifn(t=9,y,0) <= 8; quit; proc print data=o; run;

LzEr23 · ‎06-03-2019

Thank you. I did help.

kessler · ‎02-12-2019

Typically, when there is only one dependent variable in your model you use all the instrumental variables to instrument both endogenous variables; however, if you want to limit which instrumental variables are used to correct for the endogeneity of each endogenous explanatory variable you could use the following: proc tmodel data=yourdata; eq.a = y - ka*a; eq.p = y - kp*p; fit a p / ols 2sls; instruments (a, b c d) (p, q r s); quit;

rtrt · ‎07-18-2018

Ok, I will try that - thank you!

kessler · ‎09-20-2017

Although the subject of your posting is "Panel data" I don't see how the panel structure of your data is spcified in your PROC MODEL estimation. PROC MODEL does not have built in facilities to handle panel data for this model. The large standard errors may be a symptom of collinearity in your data for this model. The message about convergence indicates that the data may underspecify the parameters in your model as well. To know for sure what's causing these results it would be necessary to examine the REGION data set you are using.

Sinistrum · ‎08-31-2017

This comes in very handy, thank you very much. Keeps it all tidied up and in one place in the 'proc model' statement. I may use multiple 'tmp' as you dubbed them, with different lengths and then fit them together in the actual model equation. And thank you for elaborating on your answer and given even more hints on how I could improve further, Pu. Astonishing, how quickly this request has been solved one again.

kessler · ‎04-26-2017

If I understand your question correctly you would just need to model 'a' using one or more parameters that are also estimated. In the original PROC MODEL example 'p' plays the role of this parameter.

kessler · ‎01-27-2017

PROC MODEL supports a variety of estimation methods which can be specified using options on the FIT statement. Most of the estimation methods do not use a log likelihood function in their formulation. However, for the FIML method, a log likelhood function is evaluated and is reported in the Estimation Summary Statistics table. You can read more about the FIML method used in PROC MODEL here: http://documentation.sas.com/?docsetId=etsug&docsetVersion=14.2&docsetTarget=etsug_model_sect070.htm&locale=en#etsug.model.modfiml A log likelihood value is also used and reported when you specify your own likelihood using an 'errormodel ~ general' statement, or when you estimate a model using the t distribution.

SteveDenham · ‎09-26-2016

Sweet answer to use FCMP to come up with a Dirac function for a pulse input. That's something I wish I had known about $) years ago (of course, there was no PROC MODEL then, let alone PROC FCMP). Steve Denham

kessler · ‎09-02-2016

1) In general, the convergence criteria which you set with the CONVERGE= option used to stop the iteration process is R. This is typically how PROC MODEL determines convergence; however, convergence can also be satisified if the objective function value falls below the value of the SINGULAR= option. 2) RPC will be missing when there is no previous iteration available to compute a value. This could happen, for instance, when the initial parameter estimates meet the problem's convergence criteria. If you're seeing a missing RPC in other cases I'd be intersted in seeing your program.

kessler · ‎06-24-2016

PROC MODEL does not have a built in option to do the reweighting; however, you could specify two FIT statements to perform the reweighting yourself: proc model data=temp; fit y / outactual out=o; y = b0+b1*x1+b2*x2; outvar resid.y; fit y / data=o; h.y = resid_y**2; quit;

Online Status	Offline
Date Last Visited	‎09-06-2022 11:51 PM

Re: proc model : vector of restrictions

Re: proc model optimization

Re: proc model : vector of restrictions

Re: two stage least squares coding problem

Re: Multiple GMM Regressions by PROC MODEL

Re: Multiple GMM Regressions by PROC MODEL

Re: Multiple GMM Regressions by PROC MODEL

Re: proc model - conditional resrict

Re: GMM moment conditions for time series?

Re: 2SLS regression - multiple endogenous variable

Re: proc model : vector of restrictions

Re: Abbreviation of long sums in proc model

Re: Proc Model log likelihood

Re: Parameter estimation with proc model and ODEs

Re: How to do feasible generalized least square (FGLS) in SAS?

Re: proc model : vector of restrictions

Re: proc model optimization

Re: two stage least squares coding problem

Re: Multiple GMM Regressions by PROC MODEL

Re: proc model - conditional resrict

Re: GMM moment conditions for time series?

Re: 2SLS regression - multiple endogenous variable

Re: SUR (seemingly unrelated regression) for panel data in SAS

Re: Panel data

Re: Abbreviation of long sums in proc model

Re: Outputting 'state prediction' time series with Proc Model Switchin...

Re: Proc Model log likelihood

Re: Parameter estimation with proc model and ODEs

Re: Questions about convergence in Proc Model

Re: How to do feasible generalized least square (FGLS) in SAS?