How to invoke the Weka j48 and BayesNet result ?

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

How to invoke the Weka j48 and BayesNet result ?

Charles Zhou
Thanks a lot.And I'm sorry for the last post.

Hello everyone,I have a question here.

How to invoke the Weka j48 and BayesNet result in a b/s java application.Especially their tree and graph result.

  Regards!
 Charles Zhou
_______________________________________________
Wekalist mailing list
[hidden email]
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
Reply | Threaded
Open this post in threaded view
|

How to increase memory size

Rajni jain

Hello all

I am using weka for tree classifier for large dataset. During building of classifier in GUI, I faced out of memory error. I tried using

java -mx 100000000 -oss 100000000 from the command prompt

However I got the message

 invalid maximum heap size: -Xmx

could not create java virtual engine.

I am not a java expert. Pl. give me soem clues regarding valid values for increasing teh heap size and the stacksize.

Thanks


Rajni



Claim your space online! Share your world for free! Have a blast with MSN Spaces!
_______________________________________________
Wekalist mailing list
[hidden email]
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
Reply | Threaded
Open this post in threaded view
|

Re: How to increase memory size

Peter Reutemann
> I am using weka for tree classifier for large dataset. During building
> of classifier in GUI, I faced out of memory error. I tried using
>
> java -mx 100000000 -oss 100000000 from the command prompt
>
> However I got the message
>
>  invalid maximum heap size: -Xmx
>
> could not create java virtual engine.

You get the error message because you have a blank between the "-mx" and
the amount of memory you want to use. I know that Java is a little bit
inconsistent with the way the parameters are set...

Try this instead ("..." denotes your classpath and class you're
calling), which will start the JVM with 100MB of heap size (the "m"
makes it easier to read than those many 0s! ;-)):
    java -Xmx100m ...

Cheers, Peter

--
Peter Reutemann, Dept. of Computer Science, University of Waikato
Phone +64 (7) 838-4466 Ext. 8766

_______________________________________________
Wekalist mailing list
[hidden email]
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
Reply | Threaded
Open this post in threaded view
|

Re: How to increase memory size

Daniel Stephan
In reply to this post by Rajni jain
Those parameters may be different, if you use another VM than I do, but
for the SUN VM under Windows they should be like this:

java -Xmx100m -Xss128k
(*m is a shortcut for megabytes, while *k is obviously short for kilobytes)

Please consider that the Xss-value is PER THREAD. Means: if you make it
1MB big and your application starts 100 threads, your program will use
100MB just for thread stacks. It is generelly advised to make it as
small as possible, and test often to find a good value. I think the
default is 64k.
You can also find a good value for -Xmx by testing. Don't make it too
big, just make it big enough. If too much swapping occurs, buy more RAM :-).

(Please also consider that I don't know the -oss parameter, I was only
guessing that what you want is the -Xss parameter, which configures the
thread stack size.)

Other VMs on different platforms and/or by different vendors might use
slightly different parameters.
java -X should print detailed help to the console.

Good luck,
Daniel


Rajni jain schrieb:

> Hello all
>
> I am using weka for tree classifier for large dataset. During building
> of classifier in GUI, I faced out of memory error. I tried using
>
> java -mx 100000000 -oss 100000000 from the command prompt
>
> However I got the message
>
>  invalid maximum heap size: -Xmx
>
> could not create java virtual engine.
>
> I am not a java expert. Pl. give me soem clues regarding valid values
> for increasing teh heap size and the stacksize.
>
> Thanks
>
>
> Rajni
>
>
> ------------------------------------------------------------------------
> Claim your space online! Share your world for free!
> <http://g.msn.com/8HMAENIN/2749??PS=47575> Have a blast with MSN Spaces!
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Wekalist mailing list
>[hidden email]
>https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
>  
>


_______________________________________________
Wekalist mailing list
[hidden email]
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist