Repository /Rseslib/rseslib-3.0.1.jar:rseslib.example.TrainAndTest

/*
 * $RCSfile: TrainAndTest.java,v $
 * $Revision: 1.1 $
 * $Date: 2007/07/07 10:29:35 $
 * $Author: wojna $
 * 
 * Copyright (C) 2002 - 2007 Logic Group, Institute of Mathematics, Warsaw University
 * 
 *  This file is part of Rseslib.
 *
 *  Rseslib 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.
 *
 *  Rseslib 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 <http://www.gnu.org/licenses/>.
 */


package rseslib.example;

import java.io.File;
import java.util.ArrayList;
import java.util.Map;

import rseslib.processing.classification.ClassifierSet;
import rseslib.processing.classification.TestResult;
import rseslib.structure.attribute.ArrayHeader;
import rseslib.structure.attribute.Header;
import rseslib.structure.attribute.formats.DataFormatRecognizer;
import rseslib.structure.data.DoubleData;
import rseslib.structure.table.ArrayListDoubleDataTable;
import rseslib.structure.table.DoubleDataTable;
import rseslib.system.Report;
import rseslib.system.output.FileOutput;
import rseslib.system.output.StandardErrorOutput;
import rseslib.system.progress.EmptyProgress;
import rseslib.system.progress.StdOutProgress;

/**
 * Single test of classifiers on data
 * split into a training and a test table.
 * If only one data file is provided,
 * the method splits data randomly
 * with the ratio of the training data size
 * to the test data size 2 to 1.
 *
 * @author      Arkadiusz Wojna
 */
public class TrainAndTest
{
	/**
     * Main testing method.
     *
     * @param args Parameters of the method: path to training data,
     *             path to test data file (optionally) and path to data header.
     * @throws Exception when an error occurs.
     */
    public static void main(String[] args) throws Exception
    {
        // utworzenie pliku logowania
        if (args.length <= 0) throw new IllegalArgumentException("For rseslib format 2 or 3 arguments should be provided and for rses format 1 or 2 arguments should be provided");
        String resultDirectory = "";
        int lastSep = args[0].lastIndexOf('/');
        if (args[0].lastIndexOf('\\') > lastSep) lastSep = args[0].lastIndexOf('\\');
        if (lastSep >= 0) resultDirectory = args[0].substring(0, lastSep)+File.separator;
        resultDirectory = resultDirectory+"results"+File.separator;
        File resultDirFile = new File(resultDirectory);
        if (!resultDirFile.exists()) resultDirFile.mkdirs();
        String dataName = args[0].substring(lastSep+1);
        int dotIndex = dataName.lastIndexOf('.');
        if (dotIndex >= 0) dataName = dataName.substring(0, dotIndex);
        Report.addErrorOutput(new StandardErrorOutput());
        Report.addInfoOutput(new FileOutput(new File(resultDirectory+dataName+".txt")));

        // analiza argumentow i utworzenie naglowka
        DataFormatRecognizer rec = new DataFormatRecognizer();
        File dataFile = new File(args[0]);
        Header dataHeader = null;
        boolean oneFile = true;
        if (rec.isRses2xFormat(dataFile))
        {
            if (args.length > 2) throw new IllegalArgumentException("For rses format 1 or 2 arguments should be provided");
            dataHeader = new ArrayHeader(dataFile);
            if (args.length==2) oneFile = false;
        }
        else
        {
            if (args.length != 2 && args.length != 3) throw new IllegalArgumentException("For rseslib format 2 or 3 arguments should be provided");
            dataHeader = new ArrayHeader(new File(args[args.length-1]));
            if (args.length==3) oneFile = false;
        }

        // wczytanie tabel i podzial na treningowa i testowa czesc w przypadku pojedynczej tabeli
        DoubleDataTable trainTable = new ArrayListDoubleDataTable(dataFile, dataHeader, new EmptyProgress());
        DoubleDataTable testTable;
        if (oneFile)
        {
            ArrayList<DoubleData>[] parts = trainTable.randomSplit(2, 1);
            trainTable = new ArrayListDoubleDataTable(parts[0]);
            testTable = new ArrayListDoubleDataTable(parts[1]);
        }
        else
        {
            testTable = new ArrayListDoubleDataTable(new File(args[1]), dataHeader, new EmptyProgress());
        }

        ClassifierSet tester = new RseslibClassifiers(null).getClassifierSet();

        // uczenie klasyfikatorow na tabeli treningowej
        if (oneFile) Report.displaynl(args[0]+" (training part)");
        else Report.displaynl(args[0]);
        Report.displaynl(trainTable);
        tester.train(trainTable, new StdOutProgress());
        Report.displaynl();

        // klasyfikacja tabeli testowej
        if (oneFile) Report.displaynl(args[0]+" (testing part)");
        else Report.displaynl(args[1]);
        Report.displaynl(testTable);
        Map<String,TestResult> results = tester.classify(testTable, new StdOutProgress());
        Report.displaynl();

        // wypisanie wynikow
        Report.displayMapWithMultiLines("Classification results for test table", results);
        Report.close();
    }
}