imperial.modaclouds.monitoring.sda.weka.TimeSeriesForecasting.java Source code

Introduction

Here is the source code for imperial.modaclouds.monitoring.sda.weka.TimeSeriesForecasting.java

Source

/**
 * Copyright (c) 2012-2013, Imperial College London, developed under the MODAClouds, FP7 ICT Project, grant agreement n 318484
 * All rights reserved.
 * 
 *  Contact: imperial <weikun.wang11@imperial.ac.uk>
 *
 *    Licensed under the GNU GENERAL PUBLIC LICENSE, Version 3 (the "License");
 *    you may not use this file except in compliance with the License.
 *    You may obtain a copy of the License at
 *
 *        http://http://www.gnu.org/licenses/gpl.html
 *
 *    Unless required by applicable law or agreed to in writing, software
 *    distributed under the License is distributed on an "AS IS" BASIS,
 *    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 *    See the License for the specific language governing permissions and
 *    limitations under the License.
 */
package imperial.modaclouds.monitoring.sda.weka;

import imperial.modaclouds.monitoring.data_retriever.Client_Server;
import imperial.modaclouds.monitoring.data_retriever.ValueSet;

import java.io.*;

import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.Set;

import org.apache.commons.lang3.StringUtils;

import weka.core.Instances;
import weka.classifiers.functions.GaussianProcesses;
import weka.classifiers.functions.LinearRegression;
import weka.classifiers.functions.SMOreg;
import weka.classifiers.evaluation.NumericPrediction;
import weka.classifiers.timeseries.WekaForecaster;

public class TimeSeriesForecasting {

    public double forecast(String targetResource, String targetMetric, Map<String, String> parameters) {

        ArrayList<ArrayList<String>> timestamps = new ArrayList<ArrayList<String>>();
        ArrayList<ArrayList<String>> data = new ArrayList<ArrayList<String>>();

        ValueSet targetData = Client_Server.obtainData(targetResource, targetMetric);

        if (targetData == null) {
            System.out.println("No data received");
            return -1;
        }

        timestamps.add(targetData.getTimestamps());
        data.add(targetData.getValues());

        ArrayList<String> metricName = new ArrayList<String>();
        metricName.add("time");
        metricName.add(targetResource + targetMetric);

        int predictionStep = 0;
        String otherTarget1 = null;
        String otherTarget2 = null;
        String otherMetric1 = null;
        String otherMetric2 = null;
        String method = null;

        predictionStep = Integer.valueOf(parameters.get("predictionStep"));
        otherTarget1 = parameters.get("otherTarget1");
        otherTarget2 = parameters.get("otherTarget2");
        otherMetric1 = parameters.get("otherMetric1");
        otherMetric2 = parameters.get("otherMetric2");
        method = parameters.get("method");

        if (otherTarget1 != null) {
            ValueSet temp = Client_Server.obtainData(otherTarget1, otherMetric1);
            timestamps.add(temp.getTimestamps());
            data.add(temp.getValues());
            metricName.add(otherTarget1 + otherMetric1);
        }

        if (otherTarget2 != null) {
            ValueSet temp = Client_Server.obtainData(otherTarget2, otherMetric2);
            timestamps.add(temp.getTimestamps());
            data.add(temp.getValues());
            metricName.add(otherTarget2 + otherMetric2);
        }

        String fileName = metricName.get(1) + "Forecasting.arff";
        CreateArff.create(timestamps, data, metricName, fileName);

        double[][] value = compute(fileName, method, metricName.get(1), predictionStep);

        return value[predictionStep - 1][0];
    }

    /**
     * This method calculates the target metrics using correlation algorithm with the weka library.
     * 
     * @param arffFile   the arff file for weka
     * @param method   the time series forecasting algorithm for weka
     * @param target   the target metric
     * @param predictionStep_str   the prediction step in future
     * @return   the value of the target metric
     */
    public double[][] compute(String arffFile, String method, String target, int predictionStep) {

        double[][] result = null;

        int nbTargets = StringUtils.countMatches(target, ",") + 1;

        try {

            Instances data = new Instances(new BufferedReader(new FileReader(arffFile)));

            WekaForecaster forecaster = new WekaForecaster();

            forecaster.setFieldsToForecast(target);

            switch (method) {
            case "LinearRegression":
                forecaster.setBaseForecaster(new LinearRegression());
                break;
            case "GaussianProcess":
                forecaster.setBaseForecaster(new GaussianProcesses());
                break;
            case "SMOreg":
                forecaster.setBaseForecaster(new SMOreg());
                break;
            }

            forecaster.buildForecaster(data, System.out);

            forecaster.primeForecaster(data);

            List<List<NumericPrediction>> forecast = forecaster.forecast(predictionStep, System.out);

            result = new double[predictionStep][nbTargets];

            for (int i = 0; i < predictionStep; i++) {
                List<NumericPrediction> predsAtStep = forecast.get(i);
                for (int j = 0; j < nbTargets; j++) {
                    NumericPrediction predForTarget = predsAtStep.get(j);
                    result[i][j] = predForTarget.predicted();
                    //System.out.print("" + predForTarget.predicted() + " ");
                }
                //System.out.println();
            }
        } catch (Exception ex) {
            ex.printStackTrace();
        }
        return result;
    }
}