hello
i have problem when i learning structure with Bayesian Search
i have data 25 column and 400 line = 25*400
i get erreur vector<T> too long
help me please
BayesianSearch problem
-
- Site Admin
- Posts: 61
- Joined: Mon Nov 06, 2017 6:41 pm
Re: BayesianSearch problem
Hello,
Could you show part of code causing that error?
Could you show part of code causing that error?
-
- Posts: 15
- Joined: Mon Mar 12, 2018 5:32 pm
Re: BayesianSearch problem
data.readFile("new 2.txt");
BayesianSearch b= new BayesianSearch();
b.setMaxParents(24);
b.setRandSeed(25);
bkk.addForcedArc("age","class");
bkk.addForcedArc("pc","class");
bkk.addForcedArc("pcc","class");
bkk.addForcedArc("rbcc","class");
bkk.addForcedArc("wbcc","class");
bkk.addForbiddenArc(1, 3);
bkk.addForbiddenArc(1, 9);
bkk.addForbiddenArc(24, 1);
b.setBkKnowledge(bkk);
System.out.println(" learning");//is begin learning but never end learning
net= b.learn(data);//problem in this
System.out.println("end learning");
BayesianSearch b= new BayesianSearch();
b.setMaxParents(24);
b.setRandSeed(25);
bkk.addForcedArc("age","class");
bkk.addForcedArc("pc","class");
bkk.addForcedArc("pcc","class");
bkk.addForcedArc("rbcc","class");
bkk.addForcedArc("wbcc","class");
bkk.addForbiddenArc(1, 3);
bkk.addForbiddenArc(1, 9);
bkk.addForbiddenArc(24, 1);
b.setBkKnowledge(bkk);
System.out.println(" learning");//is begin learning but never end learning
net= b.learn(data);//problem in this
System.out.println("end learning");
-
- Site Admin
- Posts: 61
- Joined: Mon Nov 06, 2017 6:41 pm
Re: BayesianSearch problem
I forgot to ask about JVM error stack - it could be helpful.
Also, did you matched BKK with DataSet?
(code should be executed after bkk initialization)
Last thing - did forced and forbidden arcs aren't excluding each other?
Also, did you matched BKK with DataSet?
Code: Select all
bkk.matchData(data);
Last thing - did forced and forbidden arcs aren't excluding each other?
-
- Posts: 15
- Joined: Mon Mar 12, 2018 5:32 pm
Re: BayesianSearch problem
yes I matched BKK with DataSet
and no jvm error stack he terminated without error
when i use that code with dataset 9*100 he is run successfully
but with 24*400 he is terminated with msg vector<T> too long for 80%
and run successfully for 20%
and no jvm error stack he terminated without error
when i use that code with dataset 9*100 he is run successfully
but with 24*400 he is terminated with msg vector<T> too long for 80%
and run successfully for 20%
-
- Site Admin
- Posts: 61
- Joined: Mon Nov 06, 2017 6:41 pm
Re: BayesianSearch problem
I tried to reproduce that error with random data, but it was ineffective. I suspect the end of RAM memory, but 24x400 DataSet is not that big.
If your data is not confidential, archive your project and data into ZIP or RAR and send as attachment to support@bayesfusion.com
We will do our best to help you.
If your data is not confidential, archive your project and data into ZIP or RAR and send as attachment to support@bayesfusion.com
We will do our best to help you.