source: main/trunk/model-sites-dev/mars/src/java/WekaTrainArousalModel.java@ 34780

import java.util.Random;

import weka.core.converters.ConverterUtils.DataSource;
import weka.core.Instances;
 
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.Remove;

import weka.classifiers.Evaluation;
import weka.classifiers.trees.REPTree;

// Based on:
//   https://waikato.github.io/weka-wiki/use_weka_in_your_java_code/


// Also, for a more direct command-line approach see
//    https://waikato.github.io/weka-wiki/saving_and_loading_models/
//
// You save a trained classifier with the -d option (dumping), e.g.:
//
//   java weka.classifiers.trees.J48 -C 0.25 -M 2 -t /some/where/train.arff -d /other/place/j48.model
//
// And you can load it with -l and use it on a test set, e.g.:
//
//   java weka.classifiers.trees.J48 -l /other/place/j48.model -T /some/where/test.arff

     
class WekaTrainArousalModel
{
    // Scheme:       weka.classifiers.trees.REPTree -M 2 -V 0.001 -N 3 -S 1 -L -1 -I 0.0
    // Relation:     deam-essentia-features-arousal-valence-weka.filters.unsupervised.attribute.Remove-R472-weka.filters.unsupervised.attribute.Remove-R458-466
    // Instances:    1743
    
    public static void main(String[] args)
    {
    if (args.length != 2) {
        System.err.println("Error: incorrect number of command-line arguments");
        System.err.println("Usage: input_training_data.arff output-model.{model|ser}");
        System.exit(1);
    }
    String input_arff_filename = args[0];
    String output_model_filename = args[1];
    
        System.out.println("Training on ARFF file: " + input_arff_filename);

    try {
        DataSource data_source = new DataSource(input_arff_filename);
        Instances data_instances = data_source.getDataSet();

        // Filter out Valence
        //String[] filter_options = weka.core.Utils.splitOptions("-R 472");
        //String[] filter_options = weka.core.Utils.splitOptions("-R 472 -R 458-466");
        String[] filter_options = weka.core.Utils.splitOptions("-R 425,458-466,472"); // remove ordinal attributes and 'valance'
        Remove filter_remove = new Remove();
        filter_remove.setOptions(filter_options);
        filter_remove.setInputFormat(data_instances);
        Instances filtered_data_instances = Filter.useFilter(data_instances, filter_remove);

        // **** have removed R472, so class to predict is numAttributes()-1
        filtered_data_instances.setClassIndex(filtered_data_instances.numAttributes() - 1);
        

        // Build scheme/classifier
        REPTree classifier = new REPTree(); // scheme
        String[] reptree_options = weka.core.Utils.splitOptions("-M 2 -V 0.001 -N 3 -S 1 -L -1 -I 0.0");
        classifier.setOptions(reptree_options);

        classifier.buildClassifier(filtered_data_instances);
        System.out.println(classifier);


        Evaluation eval = new Evaluation(filtered_data_instances);
        Random rand = new Random(1);
        int folds = 10;
        eval.crossValidateModel(classifier, filtered_data_instances, folds, rand);
        System.out.println(eval.toSummaryString());

        System.out.println("Saving REPTree classifier model as: " + output_model_filename);
        weka.core.SerializationHelper.write(output_model_filename, classifier);

         
    }
    catch (Exception e) {
        e.printStackTrace();
    }
 
    }
}