Yui Hiroaki | 2 Mar 22:44 2003
Picon

remove my name

hi

how can I remove my address from the weka email list?

Best  Reagards
Yui
Richard Kirkby | 2 Mar 22:56 2003
Picon
Picon

Re: remove my name

This should be possible through the web interface at

http://list.scms.waikato.ac.nz/mailman/listinfo/wekalist

(edit options at the bottom of the page).

Anyway, you are now removed from the list Yui.

Richard
--

Yui Hiroaki wrote:
> 
> hi
> 
> how can I remove my address from the weka email list?
> 
> Best  Reagards
> Yui
> 
> _______________________________________________
> Wekalist mailing list
> Wekalist <at> list.scms.waikato.ac.nz
> http://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
Richard Kirkby | 3 Mar 02:11 2003
Picon
Picon

Re: Random number seed

WEKA makes use of the java.util.Random class to generate the numbers.
For details consult the Sun API documentation:

http://java.sun.com/j2se/1.4.1/docs/api/java/util/Random.html

Richard
--

> Shen Lixiang wrote:
> 
> Anyone knows the equations used to generate random number in Weka?
> What are the different outputs if different seeds number is chosen?
> 
> Lixiang
Ashraf Khalil | 4 Mar 08:13 2003
Picon
Picon

Unclassified instances

Hello,
I am running Id3 classifier agains a data set that has no missing class
values or attribute values but I still get some a number for
Unclassified instances in the evaluation results.
I wonder why I get unclassified instances while all the data I have (both
training and testing data) is without any unclassified instances?

Any answer will be appreciated

Ashraf
Kiri Wagstaff | 4 Mar 14:56 2003

Re: How to build training and testing files


On Tue, 4 Mar 2003, Liliana Calderon wrote:
> We are thinking about two possibilities (we don't know if they are
> correct):
> (1) In the training file, include data from users whose predictions
> won't be made.  In the testing file, include the data from users whose
> predictions will be made (this will have missing values for the items to
> predict, and real values for the items that the user has already rated)

This option is the better one.  Each item in the training data set should
have a value for whatever you are trying to predict (the "class value"),
because that is what the classifier uses to learn how to predict.
(Unsupervised methods, such as clustering algorithms, are a different
story, but that's not what you want for this problem.)

In the testing data, each item may or may not have a correct value
specified for the class value.  If you want to evaluate how accurate your
classifier is, you need to have the true class values specified.  (The
classifier won't use them when making predictions, but they will be used
to calculate how accurate the predictions are.)  If you don't have true
values for the test data, you can still apply the classifier and get
predictions, but you won't know how good they are.

You may also want to consider using cross-validation.  This approach uses
only one data set, where all items are labeled.  You can decide how many
"folds" to use; 10-fold cross validation is common.  The evaluator will
split the data set into 10 pieces, and then iteratively train the
classifier using 9/10 of the data and test it on the 10th (held-out) fold.
After doing this 10 times (once for each held-out fold), the evaluator
calculates the average accruacy.  See weka.classifiers.Evaluation for more
(Continue reading)

Rusma Mulyadi | 4 Mar 23:07 2003
Picon

JVM

Hi,
I am trying to process a dataset using weka and get an out of memory
error message. I then try the command in the tips and tricks (java
-mx100000000 -oss100000000 ...), but end up with the error message
below!

Unrecognized option: -Xoss100000000
Could not create the Java virtual machine. 

Any ideas on how to solve this?

Thanks!
Rusma
Sugato Basu | 4 Mar 23:29 2003
Picon

Re: JVM


Try giving: "java -Xms512M -Xmx1024M <classname>"

This would set the initial heap size to 512MB, and the max heapsize to
1024MB. You can change these parameter values according to your machine
specifications.

-Sugato

On Tue, 4 Mar 2003, Rusma Mulyadi wrote:

> Hi,
> I am trying to process a dataset using weka and get an out of memory
> error message. I then try the command in the tips and tricks (java
> -mx100000000 -oss100000000 ...), but end up with the error message
> below!
>
> Unrecognized option: -Xoss100000000
> Could not create the Java virtual machine.
>
> Any ideas on how to solve this?
>
> Thanks!
> Rusma
>
>
> _______________________________________________
> Wekalist mailing list
> Wekalist <at> list.scms.waikato.ac.nz
> http://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
(Continue reading)

Rusma Mulyadi | 4 Mar 23:35 2003
Picon

RE: JVM

Seems to start processing now!
Thanks alot!
Rusma

-----Original Message-----
From: Sugato Basu [mailto:sugato <at> cs.utexas.edu] 
Sent: Tuesday, March 04, 2003 3:30 PM
To: Rusma Mulyadi
Cc: wekalist <at> orc.cs.waikato.ac.nz
Subject: Re: [Wekalist] JVM

Try giving: "java -Xms512M -Xmx1024M <classname>"

This would set the initial heap size to 512MB, and the max heapsize to
1024MB. You can change these parameter values according to your machine
specifications.

-Sugato

On Tue, 4 Mar 2003, Rusma Mulyadi wrote:

> Hi,
> I am trying to process a dataset using weka and get an out of memory
> error message. I then try the command in the tips and tricks (java
> -mx100000000 -oss100000000 ...), but end up with the error message
> below!
>
> Unrecognized option: -Xoss100000000
> Could not create the Java virtual machine.
>
(Continue reading)

Kiri Wagstaff | 4 Mar 23:37 2003

Re: JVM


On Tue, 4 Mar 2003, Rusma Mulyadi wrote:
> I am trying to process a dataset using weka and get an out of memory
> error message. I then try the command in the tips and tricks (java
> -mx100000000 -oss100000000 ...), but end up with the error message
> below!
> Unrecognized option: -Xoss100000000
> Could not create the Java virtual machine.
> Any ideas on how to solve this?

Two ideas:

1. According to my java distribution (under Windows), the options should
   be formatted as -Xmx100000000 -Xss100000000.  Try that.

2. Does your machine have 95 megabytes of RAM free?  Although you're
   specifying a *maximum* heap size (mx), you're specifying a *fixed*
   stack size (ss) of just over 95 MB.  If you're using the correct
   syntax, try lowering the stack size (or skipping that option).  (Most
   of the Java and Weka memory usage should be off the heap, anyway.)

	Kiri

--
Kiri L. Wagstaff, Ph.D. (kiri.wagstaff <at> jhuapl.edu)    \|/
Science Applications Group, Space Department          -O-
The Johns Hopkins University Applied Physics Lab      /|\
Susan Frame | 5 Mar 13:06 2003
Picon
Picon

Graphic represenation of decision trees.

Hi,

I am applying ID3 to a medical data set and was wondering if there is a way to 
get Weka to output the results of the ID3 algorithm graphically so that the 
Specialist can interpret the results easily?

Thanks,

Susan

Gmane