theCoordinates

The engine.
Post Reply
Mastif
Posts: 16
Joined: Thu Apr 26, 2012 10:31 pm

theCoordinates

Post by Mastif »

hi,

Can I read a probability distribution (CPT) from a file.

I mean, instead of using
theCoordinates.UncheckedValue() = ....
theCoordinates.Next();

can I read the probability values from a file?
shooltz[BayesFusion]
Site Admin
Posts: 1457
Joined: Mon Nov 26, 2007 5:51 pm

Re: theCoordinates

Post by shooltz[BayesFusion] »

Mastif wrote:Can I read a probability distribution (CPT) from a file.
I'm not sure what exactly you want to achieve. Do you need to obtain the CPT from .xdsl file without using SMILE?
Mastif
Posts: 16
Joined: Thu Apr 26, 2012 10:31 pm

Re: theCoordinates

Post by Mastif »

I've been working in GeNIe and I created BN there. However, it was only prototype and my goal is work with SMILE.
I know the structure of my network and I need to create appropriate source code for further analysis. (setting evidence ->getting result and use it in my different calculations)
I've followed the tutorial 1 for SMILE and there I found :

// now fill in the conditional distribution for node "Forecast" using a system of
// coordinates. The probabilities are:
// P("Forecast" = Good | "Success" = Success) = 0.4
DSL_sysCoordinates theCoordinates (*theNet.GetNode(forecast)->Definition());
theCoordinates.UncheckedValue() = 0.4;
theCoordinates.Next();
Therefore, I've started wondering can I read the probability values from the file (e.g txt file) instead typing them.
I hope I am more clear now.
Thanks
shooltz[BayesFusion]
Site Admin
Posts: 1457
Joined: Mon Nov 26, 2007 5:51 pm

Re: theCoordinates

Post by shooltz[BayesFusion] »

Mastif wrote:Therefore, I've started wondering can I read the probability values from the file (e.g txt file) instead typing them.
If you want to programatically fill the CPTs, then it's up to your program to read the input data. SMILE does not include support for such feature.
Mastif
Posts: 16
Joined: Thu Apr 26, 2012 10:31 pm

Re: theCoordinates

Post by Mastif »

many thanks !!
Post Reply