BookmarkSubscribeRSS Feed
jlkmkl
Calcite | Level 5

Hello,

I'm using FS to forecast demand of energy. I've got a few questions... Can anybody help me?

  • The final models present "high" p and q values (for example greater than 2)... Is it possible to limit this values? What's the best way to do it?
  • I've also seen the p-values of the final models and thy're also "high". Is it possible to select the "best" model considering a specified cut-off p-value?

Thank you very much

Javier

5 REPLIES 5
ballardw
Super User

Are you saying you have a p-value greater than 2?! P-values should have a range of 0-1.

And for "high" p-value do you mean 0.8999 or 0.0001?

And "best" really should be considered in light of other items than just p-value. Suppose the cost of implementing model A with an associated p-value of 0.015 was half that of Model B with a p-value of 0.013. Model B may be "more significant" (not really and I don't want to get into that discussion again...) but in a decision making mode I suspect that Model A would often be the "better" model for practical purposes.

jlkmkl
Calcite | Level 5

No¡, of course i know which is the p-value range... I refer to the possibility of controlling the maximum order of the AR ( p & P) and MA ( q &Q ) parameters in SAS Forecast studio .

In reference to the other issue I found ARIMA models, select FS as champions , with non-significant parameters ( ie> 0.05 ) AR / MA and input variables ... i need to select statistically valid models, this concerns me as I want to automate the forecast process

udo_sas
SAS Employee

Hello -

This thread might be of interest: https://communities.sas.com/message/196164#196164

Thanks,

Udo

jlkmkl
Calcite | Level 5

Thank you Udo ,

TBC, Sas FS does NOT allow :

  • limit Pand Q orders.
  • Select only final models with all significative parameters,... to do this, could I usethe option "Refine parameters".The user guide is not very clear.Could someone especify exactly what exactlymakesthis option?

FS_Prop.png

Thanks in advance¡

Javier

udo_sas
SAS Employee

Hello -

This option controls the "SIGLEVEL" option of the ARIMA statement in the HPFDIAGNOSE procedure. It specifies the significance level to use as a cutoff value to decide the AR and MA orders. More details about this process can be found in SAS Forecast Server Procedures User’s Guide: check out the details section of chapter 5: the HPFDIAGNOSE procedure.

Thanks,

Udo

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.

Discussion stats
  • 5 replies
  • 1363 views
  • 2 likes
  • 3 in conversation