source: src/main/java/agents/uk/ac/soton/ecs/gp4j/util/DataLearner.java

package agents.uk.ac.soton.ecs.gp4j.util;

import java.io.BufferedReader;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
import java.util.Locale;
import java.util.Map;

import agents.Jama.Matrix;
import agents.uk.ac.soton.ecs.gp4j.bmc.BasicPrior;
import agents.uk.ac.soton.ecs.gp4j.bmc.GaussianProcessRegressionBMC;
import agents.uk.ac.soton.ecs.gp4j.gp.covariancefunctions.CovarianceFunctionFactory;

public class DataLearner {
        public static void main(String[] args) throws IOException {
                BufferedReader reader = new BufferedReader(new FileReader(new File(
                                "/mnt/data/berkeley-dataset/sensor1_times_0.1.txt")));
                Matrix trainX = Matrix.read(reader);

                reader = new BufferedReader(new FileReader(new File(
                                "/mnt/data/berkeley-dataset/sensor1_temps_0.1.txt")));
                Matrix trainY = Matrix.read(reader);

                GaussianProcessRegressionBMC regression = new GaussianProcessRegressionBMC();
                regression.setCovarianceFunction(CovarianceFunctionFactory
                                .getNoisySquaredExponentialARDCovarianceFunction());
                // .getNoisy2DTimeSquaredExponentialCovarianceFunction());

                // BasicPrior lengthScalePrior = new BasicPrior(5, 5.0, 0.5);
                BasicPrior timeScalePrior = new BasicPrior(10, 5000, 0.15);
                BasicPrior signalVariance = new BasicPrior(10, 10, 0.15);
                BasicPrior noise = new BasicPrior(1, 0.4, 0.3);

                regression.setPriors(timeScalePrior, signalVariance, noise);

                // int batchSize = 1;

                regression.updateRegression(trainX, trainY);

                printHyperParamWeights(regression.getHyperParameterWeights(), 1);

                // for (int i = 0; i < trainX.getRowDimension(); i++) {
                //
                // System.out.println(i);
                //
                // regression.updateRegression(trainX.getMatrix(i, i, 0, 0), trainY
                // .getMatrix(i, i, 0, 0));
                //
                // // if (regression.getTrainingSampleCount() > 120)
                // // regression.downdateRegression();
                //
                // printHyperParamWeights(regression.getHyperParameterWeights(), i);
                // }
        }

        private static void printHyperParamWeights(
                        Map<Double[], Double> hyperParameterWeights, int round)
                        throws IOException {

                StringBuffer buffer = new StringBuffer();

                for (Double[] hyper : hyperParameterWeights.keySet()) {
                        Double weight = hyperParameterWeights.get(hyper);

                        buffer.append(String.format(Locale.US, " %6d", round));

                        for (int j = 0; j < hyper.length; j++) {
                                buffer.append(String.format(Locale.US, " %15.5f", hyper[j]));
                        }

                        buffer.append(String.format(Locale.US, " %15.5f\n", weight));
                }

                // FileUtils.writeStringToFile(new File("params", "params-" + round
                // + ".txt"), buffer.toString());
        }
}