Messages

1

Java / Re: how to get the final CPT values

« on: June 29, 2009, 11:22:24  »

what can be the EXPRESSION for the distribution of probabilities for the above scenario where

child node has two or more parents, and its probability depends on them. For less number of parents, its easier to adapt CPT values explicitly but for more parents its not feasible as the CPT size grows exponentially.

2

Java / Re: how to get the final CPT values

« on: June 27, 2009, 17:26:10  »

(in connection to my initial post in the thread)

I have to generate the CPT for a v.simple scenario using Java API, of which I am not used to.

I have this (sample) problem of calculating the conditional probability table of the scenario that "Its raining"

it has two parent conditions,
either the ground is wet or not wet, and
either clouds are present or not present.

Now, the CPT table will contain 8 values to be filled in for the case of "is it raining or not".

(as it seems, the W and C are independent )

So, to find, P(R|W,C) = P(W|R,C) P(R|C) / P(W|C)

But, this formula doesnt solve my problem, I need to assume the 8 values that if,

- ground is Wet and Clouds are present than P(raining) is 1 else 0
- ground is Wet and Clouds are not present than P(raining) is 0.8 else
0.2
- ground is not Wet and Clouds are present than P(raining) is 0.2 else 0.8
- ground is not Wet and Clouds are not present than P(raining) is 0 else 1


Can I do it via some formula/API ? If even I have to put the CPT (conditional probability table) values myself, but what to do when I have to see the effect of "Likelihood" of some event (in this case the likelihood of W and C), i.e. lets say the likelihood of W is 70% and likelihood of C is 50%. How does it affect the CPT values ??

Any formula or way to find CPT with likelihood condition ?

Thanks in advance.

3

General Discussion / Re: modelling a simple BN

« on: June 15, 2009, 17:48:48  »

oh yes... thanks

4

General Discussion / Re: modelling a simple BN

« on: June 15, 2009, 13:22:45  »

thanks, that solved the problem, to get P(max), I propagated with MAX like below

domain.propagate (Domain.H_EQUILIBRIUM_MAX,
                Domain.H_EVIDENCE_MODE_NORMAL);

and than later to get the max P(most likely configuration, evidence) i used "domain.getSignificanceLevel()"

5

General Discussion / Re: modelling a simple BN

« on: June 15, 2009, 12:15:12  »

thanks for all the help lately.....i just came back to working on hugin again...

i want to know the Java API function for getting the Max-Propagation Value of Domain/Model that I have loaded as a .net file. I could not find it and i need it badly.

thanks in advance.

6

Java / how to get the final CPT values

« on: June 12, 2009, 18:09:41  »

hugin newbie..


Lets suppose I have two trees, each having 3 parents and 1 child,

I want to compare their probabilities as P1(A1|B1,C1,D1) and P2(A2|B2,C2,D2), where as the 3 parents are independent of each other.

how to do this via API?

7

Java / Re: Alternate Input methods and API samples

« on: April 08, 2009, 18:44:59  »

that solves the one of my questions....
I run the test file LoadAndPropagate.java in Test folder. But it always gives exception of loading Domain. The .net file is placed at the right location and it reads it but fails to excute after this line

Code: [Select]

            Domain domain = new Domain (netName + ".net", parseListener);
it says:

Exception caught: A fatal error occurred during an input or output operation.  This could be a failure to open a specified file, a failure to create a file due to wrong permissions, a failure during a write operation due to the disk running full, etc.

my 2nd question was realted to alternate input formats....instead of .net file, I want to make an xml tree to load as a network....still searching answers....

8

Java / Re: Alternate Input methods and API samples

« on: April 08, 2009, 11:59:37  »

In addition,

Do we always have to provide the .net file for java api to run it or we can create the network on the go while running behind some application. The reason i am asking this question is,

I have an application which includes some data that I will use to create my network-of-nodes. But, the data is not discrete and may change with the change of selection (of options), eventually creating the network at run time.
So, the problem is I dont have the .net file already built to do the reasoning instead, it depends on the dynamic input coming from the user-interaction via application.

9

General Discussion / What is PHPP ?

« on: April 08, 2009, 11:50:18  »

As I browsed through the forum I found this term PHPP used quite often but I could not find any abbreviation/explanation of PHPP!

10

Java / Alternate Input methods and API samples

« on: April 07, 2009, 18:02:40  »

In order to run any BN with hugin java api, one has to give as argument the .net file which has the network info. Is there any possibility to take xml based (rdf-triples) or anyother input representation formats as well ?


the examples given with hugin release are only .net/.oobn files....there should be some API realted examples as well in order to know, how to use and run.

11

HUGIN GUI Discussion / Re: Uncertain Reasoning

« on: March 24, 2009, 15:54:25  »

Using Table Generator is not an easy task....i am stuck in building expressions and completing it successfully. I tried few scenarios given as printscreen in the hugin help docs as:




but i could not made this scenario...... may be a couple of sample networks with expressions should be given with helping material to make things easier....

at one time, i tried to run the expression "Distribution(C1,C2)" over C3 (C1 and C2 are the parents of C3, all are interval nodes), it gave me some CPT values which appeared ok. But, when I tried again, the same thing, it doesnt work.

12

General Discussion / modelling a simple BN

« on: March 17, 2009, 14:38:47  »

what set of tutorials and exmaples i have to follow to build a BN for uncertain reasoning....i want to achieve the scenario, where we passed two or three conditions out of eight or nine conditions, and than the final inference node should show the probability of it being happening (either from max-propagation or sum propagation) IMAGE Attached


and BTW I need to clear some confusions between
Sum propagation and Max Propagation

13

HUGIN GUI Discussion / Uncertain Reasoning

« on: March 13, 2009, 18:05:36  »

Hello All,

I am a newbie in HUGIN GUI. I am studying uncertain reasoning in order to give suggestions or persuasions to the user. Lets consider I want to simulate the scenario similar to one given in examples as well.

e.g. given properties of being "Fit" (healthy) as,

"Eat Vegetables" , "Do exercise", "Sleep Well" ,etc.... these are the nodes on which I have to reason that out of 10 possible properties of being "Fit", the person only follows 4, so the percentage of being fit will be 40% but

- How should I make its BN?
- Are the CPT values same as probabilistic weight assigned explicitly to some node? Because I am not sure.


Apart from that, for uncertain reasoning, I should not have to give the CPT values for the expected value .e.g in the above case of being "Fit". IF that is the case than if there are 20 properties of being fit, than I had to enter more than hundred values in the CPT of node "Fit"?

- How to predict without entering the CPT values?

One more question regarding the same above scenario is:
If I have 2 expected values like "Fit" and "Fit2" from the given 10 properties, whereas, Fit precentage will be if 6/10 properties are matched than we will say its 60% Fit and if the rest of them will be assigned to "Fit2" i.e. 40%. How to achieve this?



I am quite novice in it. Sorry for that. Still learning.


thanks alot