Load and save Weka Model using Java API?

Question

I have my model on my hard drive at d:\MultiNomial.model. That model can be run correctly from weka. The model was built to classify a text using StringToVector as a filter. I am using java to load that model using Weka API. This is my source code

import weka.core.*;
import weka.classifiers.bayes.NaiveBayesMultinomial;
import weka.core.converters.ConverterUtils.DataSource;
public class Classifier {

    public static void main(String[] args) throws Exception {
        // TODO Auto-generated method stub

        NaiveBayesMultinomial    NBM =(NaiveBayesMultinomial) weka.core.SerializationHelper.read("D:/MultiNomial.model");
        DataSource source= new DataSource(   "D:/test.arff");
        Instances Testset=source.getDataSet() ;
        Testset.setClassIndex(Testset.numAttributes()-1);

        Instance newInstance=Testset.instance(0);
        double PredictVal=NBM.classifyInstance(newInstance);
        System.out.println(PredictVal);

    }

}

but I had the following error when I tried to run it from Eclipse:

Exception in thread "main" java.lang.ClassCastException: weka.classifiers.meta.FilteredClassifier cannot be cast to weka.classifiers.bayes.NaiveBayesMultinomial

what is the issue?

In Weka I used FilteredClassifier -> Unsupervised-> Attribute->StringToVector. Then I chose NaiveBayesMultinomial and I saved my model in my drive. Now When I used Java, I got the error : weka.classifiers.meta.FilteredClassifier cannot be cast to weka.classifiers.bayes.NaiveBayesMultinomial.

I think there must be a way to tell the code how to reverse StringToVector then use :

NaiveBayesMultinomial  NBM =(NaiveBayesMultinomial) weka.core.SerializationHelper.read("D:/MultiNomial.model")

As the exception states, the serialized classifier is not of the type you're trying to cast it. — qqilihq
– qqilihq, Commented Feb 20, 2016 at 18:58
I am sure they are the same model classifier algorithm. I also created another model J48 and casted to J48 and have the same error !! weka.classifiers.meta.FilteredClassifier cannot be cast to weka.classifiers.trees.J48 — Lylia John
– Lylia John, Commented Feb 20, 2016 at 19:05
Is the problem because I use FilteredClassifier "StringToVector" and the code does not support how can be done in reverse mode, before applying the Classifier algorithm ? — Lylia John
– Lylia John, Commented Feb 20, 2016 at 20:02

김진헌 · Accepted Answer · 2016-02-29 16:20:51Z

1

try to

NaiveBayesMultinomial NBM = new NaiveBayesMultinomial();
NBM = (NaiveBayesMultinomial)weka.core.SerializationHelper.read("D:/MultiNomial.model");

answered Feb 29, 2016 at 16:20

김진헌

517 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Load and save Weka Model using Java API?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related