could anyone suggest how to get only the class attribute values in the consequent of a rule...
this implies the class attribute should not appear in the antecedent of the rule....
i will appreciate your help.....
Yahoo! India Matrimony: Find your partner online.
Could some one give an example of a cost sensitive matrix that is
compatible with the latest version of weka. I have two classes and the
cost of misclassifying one is 9 times that of the other.
Also could someone show how to use naive bayes in a cost sensitive way?
I think its using the meta cost sensitive classifier but I could be wrong.
-Subramanyam B Chitti (chittis(a)cc.gatech.edu)
does someone know a method for the discretization of
continuous variables more than binar for decision
trees (i'm trying to find a ternar quarternar or more
also i'd like to know if there are different methods
in the treatement of continuous variables than C4.5
also for decision trees.
Yahoo! Mail : votre e-mail personnel et gratuit qui vous suit partout !
Créez votre Yahoo! Mail sur http://fr.benefits.yahoo.com/
Dialoguez en direct avec vos amis grâce à Yahoo! Messenger !Téléchargez Yahoo! Messenger sur http://fr.messenger.yahoo.com
I am working on a kind of supervized clustering, i.e.
a segmentation. I would like to associate to each
instance a label identifying the leaf of a
classification or model tree build.
Currently, after building a tree, given a new instance
it is possible to predict the value/class of that
instance. How can I predict, together with the value,
the leaf in which the instance falls?
Do you Yahoo!?
Win a $20,000 Career Makeover at Yahoo! HotJobs
Does anybody have a copy of the Reuters text corpus in arff (or any other
easy to analyse or convert into arff) format that I could use? I would
like to avoid having to convert the original SGML
formatted files available from the KDD archive. I have seen
messages on this in the past but no URL's to download the data from.
Computer Learning Research Centre (www.clrc.rhul.ac.uk)
Computer Science Department (www.cs.rhul.ac.uk)
Royal Holloway University of London, Egham,
Surrey, TW20 0EX, England
I downloaded the weka tool and trying to use it (under windows). My data is in a
database (MySQL DBMS) on a server which doesn't have a web interface. Usually I
have to use secure shell(only) for login to server and then connect to MySQL by
providing another password and then I can connect to my database. So, now I
searched the tutorial and archives for information and I could not find any!!
I've been using Weka for some time, the Windows version, mainly trough the
I'd need to run the same classifier several times (e.g. NaiveBayes) each
time changing only one value (like the random seed), and saving for each
iteration the results (well, just the confusion matrix would be enough).
Is there a way to do this?
I apologize if the question is trivial, but I have little experience with
java and have no idea on where to put my hands on.
Thanks for your time,
Two of the nice things about Weka are
(1)FilteredClassifier allows users to have some "extra" fields in the
data file (e.g. ID for each training example) while "ignoring" them when
building models. e.g A command to filter out example's ID (type string)
java weka.classifiers.FilteredClassifier -F
weka.filters.AttributeTypeFilter -B weka.classifiers.SMO -t train.data
(2) CVParameterSelection can perform parameter selection internally
(i.e. user don't need their own design). e.g. A command to choose best
parameter C for SMO would be
java weka.classifiers.CVParameterSelection -W weka.classifiers.SMO -P "C
1 50 5" -t train.data -T test.data
It's appealing to combine these "meta classifiers" together into a
single command. However, I had problems with this even after trying to
add "" or -- symbols as described in the "An introduction to using Weka
3.3.5 from the command line" of the "Documentation" part of Weka's main
Did anyone succeed with this before? Welcome suggestions.
Hello to all,
I've tried to use the K* weka alghorithm to solve a numeric prediction problem,
and it is seemed to me very good, while it is one of the few algorithms tha can
making numeric prediction working with both numerical and nominal attributes, so
I don't have the need to discretize the class before.
The problem is that I've seen that it's not possiblke to build a sort of
confusion matrix, and it's obvious, while it?s a numerical and not a class
problem. But I've the need to have an error-measure for each istance. I've seen
that it's possible to visualize all istances, and for each one to see the
effective value of the class and that of the predicted class.
I have for you some questions:
1) How can I compute classifier performances in this case?
2) How I should set, in the options window K* options in order to have a more
3) Is weka's grafic representation of test instances "editable" in any manner?
Thanks to all
> I was wondering if there are any thoughts of writing a new updated
> version with all the new features of weka.
We are currently in the process of updating the book. All going well
this should be done by the end of the year.