perform RBF neural network using weka

[bhanu prakash]

I would like to perform RBF neural network using weka tool but I don't find RBF under classifications I am using weka version 3.8.2.

Can you please help me.

Thanks and Regards

Bhanuprakash Vattikuti

+91 8977569778

bhanu...@gmail.com

[Peter Reutemann]

> I would like to perform RBF neural network using weka tool but I don't find
> RBF under classifications I am using weka version 3.8.2.

Not familiar with that classifier, but did you install the relevant
package and started up the JVM with package support (default: no
packages get loaded)?

Cheers, Peter
--
Peter Reutemann
Dept. of Computer Science
University of Waikato, NZ
+64 (7) 858-5174
http://www.cms.waikato.ac.nz/~fracpete/
http://www.data-mining.co.nz/

[bhanu prakash]

I have added RBF using package manager and from python python code I am using packages when starting jvm.

I am encountering new type of error.

code :

import os
import sys
import traceback
import weka.core.jvm as jvm
from weka.core.classes import Random
import weka.core.converters as converters
from weka.classifiers import Classifier, Evaluation




def main(args):

    # load a dataset
    data = converters.load_any_file("KDDTrain.csv")
    data.class_is_last()

    # generate train/test split of randomized data
    train, test = data.train_test_split(80.0, Random(1))

    # build classifier
    cls = Classifier(classname="weka.classifiers.functions.RBFNetwork")
    cls.build_classifier(train)
    print(cls)

    # evaluate
    evl = Evaluation(train)
    evl.test_model(cls, test)
    print(evl.summary())


if __name__ == "__main__":
    try:
        jvm.start(system_cp=True, packages=True)
        main(sys.argv)
    except Exception as e:
        print(traceback.format_exc())
    finally:

        jvm.stop()

error:

C:\Users\bhanu\PycharmProjects\main_project\venv\Scripts\python.exe C:/Users/bhanu/PycharmProjects/main_project/multi_layer_nn.py

DEBUG:weka.core.jvm:Adding bundled jars

DEBUG:weka.core.jvm:Adding system classpath

DEBUG:weka.core.jvm:Classpath=['C:\\Users\\bhanu\\PycharmProjects\\main_project\\venv\\lib\\site-packages\\javabridge\\jars\\rhino-1.7R4.jar', 'C:\\Users\\bhanu\\PycharmProjects\\main_project\\venv\\lib\\site-packages\\javabridge\\jars\\runnablequeue.jar', 'C:\\Users\\bhanu\\PycharmProjects\\main_project\\venv\\lib\\site-packages\\javabridge\\jars\\cpython.jar', 'C:\\Users\\bhanu\\PycharmProjects\\main_project\\venv\\lib\\site-packages\\weka\\lib\\python-weka-wrapper.jar', 'C:\\Users\\bhanu\\PycharmProjects\\main_project\\venv\\lib\\site-packages\\weka\\lib\\weka.jar']

DEBUG:weka.core.jvm:MaxHeapSize=default

DEBUG:weka.core.jvm:Package support enabled

DEBUG:javabridge.jutil:Creating JVM object

DEBUG:javabridge.jutil:Signalling caller

java.lang.IllegalArgumentException: Can't normalize array. Sum is NaN.

weka.core.Utils.normalize(Utils.java:1327)

weka.core.Utils.logs2probs(Utils.java:1357)

weka.filters.unsupervised.attribute.ClusterMembership.convertInstance(ClusterMembership.java:360)

weka.filters.unsupervised.attribute.ClusterMembership.batchFinished(ClusterMembership.java:273)

weka.filters.Filter.useFilter(Filter.java:704)

weka.classifiers.functions.RBFNetwork.buildClassifier(RBFNetwork.java:203)

at weka.core.Utils.normalize(Utils.java:1327)

at weka.core.Utils.logs2probs(Utils.java:1357)

at weka.filters.unsupervised.attribute.ClusterMembership.convertInstance(ClusterMembership.java:360)

at weka.filters.unsupervised.attribute.ClusterMembership.batchFinished(ClusterMembership.java:273)

at weka.filters.Filter.useFilter(Filter.java:704)

at weka.classifiers.functions.RBFNetwork.buildClassifier(RBFNetwork.java:203)

Traceback (most recent call last):

  File "C:/Users/bhanu/PycharmProjects/main_project/multi_layer_nn.py", line 35, in <module>

    main(sys.argv)

  File "C:/Users/bhanu/PycharmProjects/main_project/multi_layer_nn.py", line 23, in main

    cls.build_classifier(train)

  File "C:\Users\bhanu\PycharmProjects\main_project\venv\lib\site-packages\weka\classifiers.py", line 83, in build_classifier

    javabridge.call(self.jobject, "buildClassifier", "(Lweka/core/Instances;)V", data.jobject)

  File "C:\Users\bhanu\PycharmProjects\main_project\venv\lib\site-packages\javabridge\jutil.py", line 887, in call

    result = fn(*nice_args)

  File "C:\Users\bhanu\PycharmProjects\main_project\venv\lib\site-packages\javabridge\jutil.py", line 854, in fn

    raise JavaException(x)

javabridge.jutil.JavaException: Can't normalize array. Sum is NaN.

Exception in thread "Thread-0" 

Process finished with exit code 0 

I request you to please help me

[Peter Reutemann]

> I have added RBF using package manager and from python python code I am using packages when starting jvm.
>
> I am encountering new type of error.
>
> code :

[...]

> java.lang.IllegalArgumentException: Can't normalize array. Sum is NaN.
> weka.core.Utils.normalize(Utils.java:1327)
> weka.core.Utils.logs2probs(Utils.java:1357)
> weka.filters.unsupervised.attribute.ClusterMembership.convertInstance(ClusterMembership.java:360)
> weka.filters.unsupervised.attribute.ClusterMembership.batchFinished(ClusterMembership.java:273)
> weka.filters.Filter.useFilter(Filter.java:704)
> weka.classifiers.functions.RBFNetwork.buildClassifier(RBFNetwork.java:203)

[..]

Once again, always test with different datasets to determine whether
your dataset is the problem. Your code works fine with the UCI
datasets bolts (numeric class) and anneal (nominal class).
I recommend, that you convert your CSV file into an ARFF file and
inspect the attribute types that it generated. Maybe there is
something odd going on.
Last but not least, use the Weka mailing list for problems with a
particular Weka algorithm. I can only provide help with the wrapper
itself.