Continuous variables - body temperature

The front end.
Post Reply
paaddor
Posts: 3
Joined: Mon Aug 25, 2014 9:38 am

Continuous variables - body temperature

Post by paaddor »

Hi

I'm looking for the way how to model continous variables in GeNIe.

In the GeNIe Help I'm reading under the title "Discrete and Continous variables":

"Continuous variables can assume an infinite number of values. An example of a continuous variable is Body temperature, assuming any value between 30 and 45 degrees Celsius."

Does somebody knows where I can find "Body temperature"?
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: Continuous variables - body temperature

Post by shooltz[BayesFusion] »

Does somebody knows where I can find "Body temperature"?
The body temperature was mentioned as an example of continuous variable (node in the network). GeNIe has no built-in "body temperature" (or any other built-in variables); users have to create nodes manually or learn the network structure from the data.
paaddor
Posts: 3
Joined: Mon Aug 25, 2014 9:38 am

Re: Continuous variables - body temperature

Post by paaddor »

Thanks alot for the quick answer. I see that "body temperature" isn't an example network.

So my question is how to model a continous variable that follows to a certain distribution. I didn't find any remark in the Help.

I could even imagine that I could pull and put a Bézier curve to model the distribution of a variable as within the attachment. Here the node "Adjusted Duration" has two parents, "Duration" and "Adjustment Factor", both continous. "Adjustment Factor" is modeled as a Bézier curve. "Duration" has also two parents, both discrete.

I'm wundering how this is realized within a Bayesian network tool.
Attachments
Unbenannt.JPG
Unbenannt.JPG (65.27 KiB) Viewed 4950 times
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: Continuous variables - body temperature

Post by shooltz[BayesFusion] »

So my question is how to model a continous variable that follows to a certain distribution.
You need to add a continuous node and set the required distribution in the "Definition" page of its properties. There are multiple probability distributions available - see the "Functions and operators" pane. At this point there's no "freehand" option with bezier representing the PDF. The closest thing would be the CustomPDF function, which accepts the list of numbers representing the point on the PDF line-segment approximation.
paaddor
Posts: 3
Joined: Mon Aug 25, 2014 9:38 am

Re: Continuous variables - body temperature

Post by paaddor »

"add a continuous node"? May be you mean an equation node? There I can choose probability distributions.

Unfortunately the GeNIe documentation doesn't tell anything about equation nodes and its integration in a BN.
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: Continuous variables - body temperature

Post by shooltz[BayesFusion] »

paaddor wrote:"add a continuous node"? May be you mean an equation node? There I can choose probability distributions.
Equation node === continuous node.
Post Reply