Search code examples
javaapiweka

Java WEKA API - StratifiedRemoveFolds


I have ~100 datasets in csv format and I'd like to apply Weka filter StratifiedRemoveFolds.
My code so far is:

CSVLoader loader = new CSVLoader();
loader.setSource(new File(datasetToBeProcessed.toString()));
Instances data = loader.getDataSet();

StratifiedRemoveFolds strRmvFolds = new StratifiedRemoveFolds();
strRmvFolds.setFold(fold);
strRmvFolds.setNumFolds(DatasetPreprocessing.numberOfFolds);
strRmvFolds.setSeed(0);
strRmvFolds.setInvertSelection(false);
strRmvFolds.setInputFormat(data);
data = StratifiedRemoveFolds.useFilter(data, strRmvFolds);

When I run this I get the following error message:

stratifyDatasetsWithWekaApi 
SEVERE: null
weka.core.UnassignedClassException: 
weka filters.supervised.instance.StratifiedRemoveFolds: 
Class attribute not set!

I've tried the following as well:

CSVLoader loader = new CSVLoader();
loader.setSource(new File(datasetToBeProcessed.toString()));
Instances data = loader.getDataSet();

String[] options = new String[4]; // initialize options
options[0] = "-S 0"; // seed number
options[1] = "-V"; // invert selection
options[2] = "-N 8"; // total number of folds
options[3] = "-F 1";  // current fold

strRmvFolds.setOptions(options);
strRmvFolds.setInputFormat(data);
data = StratifiedRemoveFolds.useFilter(data, strRmvFolds);

Error message is:

stratifyDatasetsWithWekaApi
SEVERE: null
java.lang.Exception: Illegal options: -S 0 -N 8 -F 1 

What am I doing wrong?

Thanks, Pavlos

http://weka.sourceforge.net/doc.dev/weka/filters/supervised/instance/StratifiedRemoveFolds.html


Solution

  • You need to set the class index for loaded instances. Class index is usually the last index. You can set it like :

    Instances data = loader.getDataSet();
    data.setClassIndex(data.numAttributes() - 1);