Can somebody help me with calling weka algorithms in matlab?

Hello,

I'm working on machine learning techniques and instead of using WEKA workbench, I want to use the same algorithms but integrate in Matlab. the first step, creating a classifier and classifying an unknown instances I can do it in matlab but i'm facing with problem in extracting the parameters such as number of correctly classified instances, confusion matrix and so on. here's the way how I do it:

    %%%include jar libraries to matlab environment:
    javaaddpath('C:\Program Files\MATLAB\R2011a\java\jar\weka.jar');

    %%%importing of classes
    import java.util.Enumeration
    import java.lang.String
    import weka.classifiers.Classifier
    import weka.classifiers.Evaluation
    import weka.classifiers.trees.J48
    import java.io.FileReader               
    import weka.core.Instances
    import weka.core.Utils
    import weka.core.Attribute
    import java.lang.System

    %%%Reading and configuring treaining set
    reader = javaObject('java.io.FileReader','TrainSetOfInstances.arff');
    dataset = javaObject('weka.core.Instances',reader);
    dataset.setClassIndex(dataset.numAttributes() - 1);

    %%%Reading and Configuring testing set (in case we want to perform train/test evaluation)
    reader1 = javaObject('java.io.FileReader','TestSetOfInstances.arff');
    testset = javaObject('weka.core.Instances',reader1);
    testset.setClassIndex(testset.numAttributes() - 1);

    %%%getting the parameters of an arff file
    relationName = char(dataset.relationName);
    numAttr = dataset.numAttributes;
    numInst = dataset.numInstances;

    %%%creating an object of class J48 and setup the default options
    classifier = javaObject('weka.classifiers.trees.J48');
    classifier.setOptions(weka.core.Utils.splitOptions('-c last -C 0.25 -M 2'));

    %%%string for setting the options
    v1 = String('-t');
    v2 = String('TrainSetOfInstances.arff');
    v3 = String('-x');        %%%if we want cross-validation
    v4 = String('10');        %%%with number of folds

    options = cat(1,v1,v2,v3,v4);

    %%%Creation of classifier Model
    classifier.buildClassifier(dataset);

    %%%evaluation of the Classifier with options
    weka.classifiers.Evaluation.evaluateModel(classifier,options)

    %%%creating an object containing the evaluations
    eval = weka.classifiers.Evaluation(dataset);
    eval.evaluateModel(classifier, options1);
    %%%this part works properly

    %%%printing the parameters of the evaluation
    fprintf('=== Run information ===\n\n')
    fprintf('Scheme: %s %s\n', ...
        char(classifier.getClass().getName()), ...
        char(weka.core.Utils.joinOptions(classifier.getOptions())) )
    fprintf('Relation: %s\n', char(dataset.relationName))
    fprintf('Instances: %d\n', dataset.numInstances)
    fprintf('Attributes: %d\n\n', dataset.numAttributes)
    %%%this part works properly

    fprintf('=== Classifier model ===\n\n')
    disp( char(classifier.toString()) )
    %%%this part works properly

    fprintf('=== Summary ===\n')
    disp( char(eval.toSummaryString()) )
    disp( char(eval.toClassDetailsString()) )
    disp( char(eval.toMatrixString()) )
    %%%This part does not work properly, all of the time I get zeros and dont know why!

So, my problem is, when I run the evaluation command:

    %%%evaluation of the Classifier with options
    weka.classifiers.Evaluation.evaluateModel(classifier,options)

in the command line I get the classification tree, evaluation model, confusion matrix and all others statistics that I get in weka workbench but when I try to extract them using following methods: toSummaryString(), toClassDetailsString() or toMatrixString() I get only zeros!

So, can someone tell me what I don't do or what I do wrong?

Thanks

[Asdrubal Lopez Chau]

Hi Aleksandar Petrov, maybe this can help you

clc
clear
%Load from disk
fileDataset = 'cm1.arff';
myPath = 'C:\Users\Asdrubal\Google Drive\Respaldo\DoctoradoALCPC\Doctorado ALC PC\AlcMobile\AvTh\MyPapers\Papers2014\UnderOverSampling\data\Skewed\datasetsKeel\';
javaaddpath('C:\Users\Asdrubal\Google Drive\Respaldo\DoctoradoALCPC\Doctorado ALC PC\AlcMobile\JarsForExperiments\weka.jar');
wekaOBJ = loadARFF([myPath fileDataset]);
%Transform from data into Matlab
[data, featureNames, targetNDX, stringVals, relationName] = ... 
weka2matlab(wekaOBJ,'[]');
%Create testing and training sets in matlab format (this can be improved)
[tam, dim] = size(data);
idx = randperm(tam);
testIdx = idx(1 : tam*0.3);
trainIdx = idx(tam*0.3 + 1:end);
trainSet = data(trainIdx,:);
testSet = data(testIdx,:);
%Trasnform the training and the testing sets into the Weka format
testingWeka = matlab2weka('testing', featureNames, testSet);
trainingWeka = matlab2weka('training', featureNames, trainSet);
%Now evaluate classifier
import weka.classifiers.*;
import java.util.*
wekaClassifier = javaObject('weka.classifiers.trees.J48');
wekaClassifier.buildClassifier(trainingWeka);
e = javaObject('weka.classifiers.Evaluation',trainingWeka);
myrand = Random(1);
plainText = javaObject('weka.classifiers.evaluation.output.prediction.PlainText');
buffer = javaObject('java.lang.StringBuffer');
plainText.setBuffer(buffer)
bool = javaObject('java.lang.Boolean',true);
range = javaObject('weka.core.Range','1');
array = javaArray('java.lang.Object',3);
array(1) = plainText;
array(2) = range;
array(3) = bool;
e.crossValidateModel(wekaClassifier,testingWeka,10,myrand,array)%U
e.toClassDetailsString

[alisha murugesh]

   %%%creating an object containing the evaluations
    eval = weka.classifiers.Evaluation(dataset);
    eval.evaluateModel(classifier, options1);

I tried executing the above code, and I have a doubt in the above snippet, is it supposed to be 'options1' ,should'nt it be options, however when I change it to 'options'

This is the error I get

 Java exception occurred: java.lang.Exception: Weka exception: No training file and no object input file given.

Answer 1

[Stalin Samuel]

http://in.mathworks.com/matlabcentral/fileexchange/33128-parallel-distributed-processing-of-weka-algorithms-in-matlab