I'm trying to understand why I'm getting an infinity or missing tvalue for regression analysis (GARCH). It's a very large data set. Do you have any idea?
Thanks
For this type of question, it is best to contact SAS Tech Support: https://support.sas.com/en/technicalsupport.html#contacttechnicalsupport
Best of luck to you,
Shelley
This usually happens because the denominator is zero. Or maybe you have zero degrees of freedom for an effect.
Could you please show us your output where you see this?
My code is:
proc autoreg data=work;
model y = / garch=(p=1,q=1);
hetere x z a;
run;
I just use GARCH model with other explanatory variables.
My result is in attached file.
Can you show us your code and output?
My code is:
proc autoreg data=work;
model y = / garch=(p=1,q=1);
hetere x z a;
run;
I just use GARCH model with other explanatory variables.
My result is in attached file.
EDIT (via Reeza)
The SAS System 
The AUTOREG Procedure
GARCH Estimates 

SSE 
0.01656969 
Observations 
150 
MSE 
0.0001105 
Uncond Var 
3.05834E8 
Log Likelihood 
574.554614 
Total RSquare 
. 
SBC 
1114.0348 
AIC 
1135.1092 
MAE 
0.00422601 
AICC 
1134.3205 
MAPE 
100.402247 
HQC 
1126.5474 


Normality Test 
1294.3504 


Pr > ChiSq 
<.0001 
Parameter Estimates 

Variable 
DF 
Estimate 
Standard 
t Value 
Approx 
Intercept 
1 
0.000035 
0.001136 
0.03 
0.9754 
ARCH0 
1 
1.0537E8 
0 
Infty 
<.0001 
ARCH1 
1 
0.6555 
0.1366 
4.80 
<.0001 
GARCH1 
1 
0 
0 
. 
. 
HET1 
1 
9.587E10 
0 
Infty 
<.0001 
HET2 
1 
0 
0 
. 
. 
HET3 
1 
0 
0 
. 
Can you show us the relevant portion of your output by simply copying and pasting the text right here into the {i} window? My organization will not let me download MS Office documents.
Surely. This is result for 150 observations:
The AUTOREG Procedure
GARCH Estimates  
SSE  0.01656969  Observations  150 
MSE  0.0001105  Uncond Var  3.05834E8 
Log Likelihood  574.554614  Total RSquare  . 
SBC  1114.0348  AIC  1135.1092 
MAE  0.00422601  AICC  1134.3205 
MAPE  100.402247  HQC  1126.5474 

 Normality Test  1294.3504 

 Pr > ChiSq  <.0001 
Parameter Estimates  
Variable  DF  Estimate  Standard  t Value  Approx 
Intercept  1  0.000035  0.001136  0.03  0.9754 
ARCH0  1  1.0537E8  0  Infty  <.0001 
ARCH1  1  0.6555  0.1366  4.80  <.0001 
GARCH1  1  0  0  .  . 
HET1  1  9.587E10  0  Infty  <.0001 
HET2  1  0  0  .  . 
HET3  1  0  0  . 
I'm not too familiar with these models, but your standard errors are 0. What's the distribution of those variables?
Post a Proc Means or Proc Univariate output for the original input data?
@Khaladdin wrote:
It's a very large data set. Do you have any idea?
Your regression results state that you only have 150 observations, so either you're not working with the data you think you are or something else is wrong somewhere.
[Deleted, via @Reeza]
If you want to repost this, feel free, but it didn't seem relevant to the question or topic.
@Khaladdin wrote:
Oh sorry. I have uploaded wrong documents and away from my pc now. The key point is i got infinity or missing t values.
And there are valid reasons for this occurring. Without the results we can't say if it's valid or not.
For this type of question, it is best to contact SAS Tech Support: https://support.sas.com/en/technicalsupport.html#contacttechnicalsupport
Best of luck to you,
Shelley
Don't miss out on SAS Innovate  Register now for the FREE Livestream!
Can't make it to Vegas? No problem! Watch our general sessions LIVE or ondemand starting April 17th. Hear from SAS execs, bestselling author Adam Grant, Hot Ones host Sean Evans, top tech journalist Kara Swisher, AI expert Cassie Kozyrkov, and the mindblowing dance crew iLuminate! Plus, get access to over 20 breakout sessions.
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.