ai.preferred.regression.TrainLogisticRegression Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of csvpl Show documentation
Your preferred data manipulation and analysis language for CSV.
The newest version!
/*
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see .
 */

package ai.preferred.regression;

import ai.preferred.regression.io.ARFFDataReader;
import org.kohsuke.args4j.Option;
import weka.classifiers.evaluation.Evaluation;
import weka.core.Instances;
import weka.core.SerializationHelper;

import java.io.File;
import java.io.FileOutputStream;

public class TrainLogisticRegression extends Command {

  @Option(name = "-i", aliases = {"--train"}, usage = "the path to the training data in CSV format", required = true)
  private File input;

  @Option(name = "-m", aliases = {"--model"}, usage = "the output path to the model file", required = true)
  private File model;

  @Option(name = "-h", aliases = {"--header"}, usage = "specifies if the input CSV files have headers")
  private boolean header = true;

  @Option(name = "-r", aliases = {"--ridge"}, usage = "the ridge parameter")
  private double ridge = 1.0;

  @Option(name = "-v", aliases = {"--verbose"}, usage = "verbosity level (-1 - disable, 0 - short, 1 - default, 2 - detailed)")
  private int verbose = 1;

  @Override
  protected void exec() throws Exception {
    final ARFFDataReader reader = new ARFFDataReader(input, true, header);
    final Instances data = reader.read(input);
    final weka.classifiers.functions.Logistic classifier = new weka.classifiers.functions.Logistic();
    classifier.setRidge(ridge);
    classifier.buildClassifier(data);

    final Evaluation eval = new Evaluation(data);
    eval.evaluateModel(classifier, data);
    if (verbose <= -1) {
      // output disabled
    } else if (verbose == 0) {
      System.out.println(eval.pctCorrect());
    } else if (verbose == 1) {
      System.out.println("ACCURACY[TRAINING] = " + eval.pctCorrect());
    } else {
      System.out.println();
      System.out.println("CLASS\tPRECISION\tRECALL\tF-MEASURE");
      for (int i = 0; i < data.classAttribute().numValues(); i++) {
        System.out.printf("%s\t%f\t%f\t%f", data.classAttribute().value(i), eval.precision(i), eval.recall(i), eval.fMeasure(i));
        System.out.println();
      }
    }

    SerializationHelper.write(new FileOutputStream(model), classifier);
  }

  public static void main(String[] args) {
    parseArgsAndRun(TrainLogisticRegression.class, args);
  }


}