Continuous variables & gaussian states

The front end.
Post Reply
oscarpc
Posts: 33
Joined: Mon Jun 04, 2012 1:17 am

Continuous variables & gaussian states

Post by oscarpc »

Hello all,

I have a couple of questions that might be a bit simple, but that I don't understand very well.
1. How can I use continuous variables in my DBN in Genie?
2. And second, is there any way of having a probability density function, let's say a Gaussian, defining the conditional probability between two or more nodes?

Thank you very much for your help.
Kind regards,

Oscar
marek [BayesFusion]
Site Admin
Posts: 430
Joined: Tue Dec 11, 2007 4:24 pm

Re: Continuous variables & gaussian states

Post by marek [BayesFusion] »

I'm afraid we have not ported continuous variables to dynamic models yet :-(. There are no theoretical reasons against it, so if you make a good case, we will push on it.

As far as your second question goes, I think it should be possible. If you tell me what precisely you want, I'll try to construct a model that does it.
Cheers,

Marek
Gardiola
Posts: 29
Joined: Mon Mar 11, 2013 4:38 pm

Re: Continuous variables & gaussian states

Post by Gardiola »

2. And second, is there any way of having a probability density function, let's say a Gaussian, defining the conditional probability between two or more nodes?

I think it s interesting question this one.
marek [BayesFusion]
Site Admin
Posts: 430
Joined: Tue Dec 11, 2007 4:24 pm

Re: Continuous variables & gaussian states

Post by marek [BayesFusion] »

There is. If you tell me what exactly you want, I will be glad to help :-). Cheers -- Marek
Gardiola
Posts: 29
Joined: Mon Mar 11, 2013 4:38 pm

Re: Continuous variables & gaussian states

Post by Gardiola »

For exemple, it's the distribution of the two parameters that describe a normal distribution.
marek [BayesFusion]
Site Admin
Posts: 430
Joined: Tue Dec 11, 2007 4:24 pm

Re: Continuous variables & gaussian states

Post by marek [BayesFusion] »

Try constructing a model consisting of three continuous variables (green node with a wave), call them mu, sigma, and x. Then you can use the values of the variables called mu and sigma to parametrize the distribution of x (essentially, define x=Normal(mu,sigma)). I hope this helps.
Cheers,

Marek
Gardiola
Posts: 29
Joined: Mon Mar 11, 2013 4:38 pm

Re: Continuous variables & gaussian states

Post by Gardiola »

Marek,

Your post shows my lack of the knowledge of Genie:

I don't know how to " define x=Normal(mu,sigma)".

Cheers.

Didier
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: Continuous variables & gaussian states

Post by shooltz[BayesFusion] »

Gardiola wrote:I don't know how to " define x=Normal(mu,sigma)".
Open the properties of node x, select the "Definition" page and type the equation "x=Normal(mu, sigma)". Next, press the OK button.
Gardiola
Posts: 29
Joined: Mon Mar 11, 2013 4:38 pm

Re: Continuous variables & gaussian states

Post by Gardiola »

Shooltz,

OK thanks for this.

I was just working on the learning of the structure for the the data first.

I totally ignored the decision dimensions so far.

I will start to work on this aspect from now: it looks very powerfull to make any decisions !!!

Cheers.

Didier
Bayesian
Posts: 2
Joined: Mon May 13, 2013 8:44 pm

Re: Continuous variables & gaussian states

Post by Bayesian »

How can we combine plain chance nodes with equation nodes?
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: Continuous variables & gaussian states

Post by shooltz[BayesFusion] »

Bayesian wrote:How can we combine plain chance nodes with equation nodes?
Currently you can't do that. This feature is high on the to-do list.
Post Reply