SAS Forecasting and Econometrics

Forecasting using SAS Visual Forecasting, SAS Forecast Server, SAS Econometrics and more
BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
SlutskyFan
Obsidian | Level 7
Is there a way to implement markov chains models in SAS ( for forecasting purposes using transition matrices as opposed to gibbs sampler MCMC simulations)

I suppose I could define matrices in R or via SAS IML, but I'd hope that there is a PROC that can be exploited for this.

For example see:

http://www.utdallas.edu/ospa/research/Conference%20Presentations/AIR/documents/TimeHorizon.pdf

Thanks
1 ACCEPTED SOLUTION

Accepted Solutions
AndreToman
Calcite | Level 5

There is no out-of-the-box PROC that can build Markov chain forecasts that I know of.  If it's really simple with static transition probabilities, then it could be done in IML.  Alternatively, one can build dynamic Markov models using Base SAS and SAS/STAT...you don't even have to use IML.  Your transition probabilities must be computed using logistic regression.  I build migration models for account delinquency status estimation (stress testing models for the Dodd-Frank Act) using SAS 9.3.  It's lots of coding.  If I can make my code sanitary enough, I'll try to share in the SAS communities.

Andre

View solution in original post

2 REPLIES 2
Eisa
Calcite | Level 5

From my knowledge, you could implement MC (markov chains) model in proc IML.

I attached an example of code that use SAS to implement MC models in proc IML.

Some options in proc MCMC export posterior sample for other analysis. However i not sure to fully understand your need ?

AndreToman
Calcite | Level 5

There is no out-of-the-box PROC that can build Markov chain forecasts that I know of.  If it's really simple with static transition probabilities, then it could be done in IML.  Alternatively, one can build dynamic Markov models using Base SAS and SAS/STAT...you don't even have to use IML.  Your transition probabilities must be computed using logistic regression.  I build migration models for account delinquency status estimation (stress testing models for the Dodd-Frank Act) using SAS 9.3.  It's lots of coding.  If I can make my code sanitary enough, I'll try to share in the SAS communities.

Andre