Example usage for weka.core Matrix Matrix

Introduction

In this page you can find the example usage for weka.core Matrix Matrix.

Prototype

public Matrix(Reader r) throws Exception 

Source Link

Document

Reads a matrix from a reader.

Usage

From source file:data.generation.target.utils.PrincipalComponents.java

License:Open Source License

private void buildAttributeConstructor(Instances data) throws Exception {
    m_eigenvalues = null;/*from  w ww.  ja  va 2  s  .  c om*/
    m_outputNumAtts = -1;
    m_attributeFilter = null;
    m_nominalToBinFilter = null;
    m_sumOfEigenValues = 0.0;
    m_trainInstances = new Instances(data);

    // make a copy of the training data so that we can get the class
    // column to append to the transformed data (if necessary)
    m_trainHeader = new Instances(m_trainInstances, 0);

    m_replaceMissingFilter = new ReplaceMissingValues();
    m_replaceMissingFilter.setInputFormat(m_trainInstances);
    m_trainInstances = Filter.useFilter(m_trainInstances, m_replaceMissingFilter);

    /*if (m_normalize) {
      m_normalizeFilter = new Normalize();
      m_normalizeFilter.setInputFormat(m_trainInstances);
      m_trainInstances = Filter.useFilter(m_trainInstances, m_normalizeFilter);
    } */

    m_nominalToBinFilter = new NominalToBinary();
    m_nominalToBinFilter.setInputFormat(m_trainInstances);
    m_trainInstances = Filter.useFilter(m_trainInstances, m_nominalToBinFilter);

    // delete any attributes with only one distinct value or are all missing
    Vector deleteCols = new Vector();
    for (int i = 0; i < m_trainInstances.numAttributes(); i++) {
        if (m_trainInstances.numDistinctValues(i) <= 1) {
            deleteCols.addElement(new Integer(i));
        }
    }

    if (m_trainInstances.classIndex() >= 0) {
        // get rid of the class column
        m_hasClass = true;
        m_classIndex = m_trainInstances.classIndex();
        deleteCols.addElement(new Integer(m_classIndex));
    }

    // remove columns from the data if necessary
    if (deleteCols.size() > 0) {
        m_attributeFilter = new Remove();
        int[] todelete = new int[deleteCols.size()];
        for (int i = 0; i < deleteCols.size(); i++) {
            todelete[i] = ((Integer) (deleteCols.elementAt(i))).intValue();
        }
        m_attributeFilter.setAttributeIndicesArray(todelete);
        m_attributeFilter.setInvertSelection(false);
        m_attributeFilter.setInputFormat(m_trainInstances);
        m_trainInstances = Filter.useFilter(m_trainInstances, m_attributeFilter);
    }

    // can evaluator handle the processed data ? e.g., enough attributes?
    getCapabilities().testWithFail(m_trainInstances);

    m_numInstances = m_trainInstances.numInstances();
    m_numAttribs = m_trainInstances.numAttributes();

    //fillCorrelation();
    fillCovariance();

    double[] d = new double[m_numAttribs];
    double[][] v = new double[m_numAttribs][m_numAttribs];

    Matrix corr = new Matrix(m_correlation);
    corr.eigenvalueDecomposition(v, d);
    m_eigenvectors = (double[][]) v.clone();
    m_eigenvalues = (double[]) d.clone();

    /*for (int i = 0; i < m_numAttribs; i++) {
      for (int j = 0; j < m_numAttribs; j++) {
        System.err.println(v[i][j] + " ");
      }
      System.err.println(d[i]);
    } */

    // any eigenvalues less than 0 are not worth anything --- change to 0
    for (int i = 0; i < m_eigenvalues.length; i++) {
        if (m_eigenvalues[i] < 0) {
            m_eigenvalues[i] = 0.0;
        }
    }
    m_sortedEigens = Utils.sort(m_eigenvalues);
    m_sumOfEigenValues = Utils.sum(m_eigenvalues);

    m_transformedFormat = setOutputFormat();
    if (m_transBackToOriginal) {
        m_originalSpaceFormat = setOutputFormatOriginal();

        // new ordered eigenvector matrix
        int numVectors = (m_transformedFormat.classIndex() < 0) ? m_transformedFormat.numAttributes()
                : m_transformedFormat.numAttributes() - 1;

        double[][] orderedVectors = new double[m_eigenvectors.length][numVectors + 1];

        // try converting back to the original space
        for (int i = m_numAttribs - 1; i > (m_numAttribs - numVectors - 1); i--) {
            for (int j = 0; j < m_numAttribs; j++) {
                orderedVectors[j][m_numAttribs - i] = m_eigenvectors[j][m_sortedEigens[i]];
            }
        }

        // transpose the matrix
        int nr = orderedVectors.length;
        int nc = orderedVectors[0].length;
        m_eTranspose = new double[nc][nr];
        for (int i = 0; i < nc; i++) {
            for (int j = 0; j < nr; j++) {
                m_eTranspose[i][j] = orderedVectors[j][i];
            }
        }
    }
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Hidden Activation Probability - returns P(Z|X).
 * A Bias column added (and removed) automatically.
 * @param   X_   X (without bias)/* w w w. j  a va  2 s  . c o  m*/
 * @return   P(Z|X) 
 */
public double[][] prob_Z(double X_[][]) {
    Matrix X = new Matrix(MatrixUtils.addBias(X_));
    return MatrixUtils.removeBias(prob_Z(X).getArray());
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Visible Activation Probability - returns P(X|Z).
 * A bias column is assumed to be included.
 * @param   Z   z (bias included)// www  .  java 2 s .  c o m
 * @return   P(X|Z) 
 */
public Matrix prob_X(Matrix Z) {
    Matrix X = new Matrix(MatrixUtils.sigma(Z.times(W.transpose()).getArray())); // (this is the activation function)
    MatrixUtils.fillCol(X.getArray(), 0, 1.0); // fix bias - set first col to 1.0
    return X;
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Make W matrix of dimensions d+1 and h+1 (+1 for biases).
 * Initialized from ~N(0,0.2) (seems to work better than ~N(0.0.01)) -- except biases (set to 0)
 * @param    d   number of rows    (visible units)
 * @param   h   number of columns (hidden units)
 * @param   r    for getting random rumbers
 * @return   W/*from www  .j  a v  a  2  s. c o m*/
 */
public static Matrix makeW(int d, int h, Random r) {
    double W_[][] = MatrixUtils.multiply(MatrixUtils.randn(d + 1, h + 1, r), 0.20); // ~ N(0.0,0.01)
    MatrixUtils.fillRow(W_, 0, 0.0); // set the first row to 0 for bias
    MatrixUtils.fillCol(W_, 0, 0.0); // set the first col to 0 for bias
    return new Matrix(W_);
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Update - On raw data (with no bias column)
 * @param   X_    raw double[][] data (with no bias column)
 *///from  w w  w.  ja  va  2 s .c  om
public void update(double X_[][]) {
    Matrix X = new Matrix(MatrixUtils.addBias(X_));
    update(X);
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Update - On raw data (with no bias column)
 * @param   x_    raw double[] data (with no bias column)
 * @param   s   multiply the gradient by this scalar
 *///  w  ww  . j  a  v  a  2s.com
public void update(double x_[], double s) {
    Matrix X = new Matrix(MatrixUtils.addBias(new double[][] { x_ }));
    update(X, s);
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Train - Setup and train the RBM on X, over m_E epochs.
 * @param   X_   X//w  ww.  j a va2  s . c o m
 * @return   the error (@TODO unnecessary)
 */
public double train(double X_[][]) throws Exception {

    initWeights(X_);

    Matrix X = new Matrix(MatrixUtils.addBias(X_));

    double _error = Double.MAX_VALUE; // prev error , necessary only when using m_V

    // TRAIN FOR m_E EPOCHS.

    for (int e = 0; e < m_E; e++) {

        // BREAK OUT IF THE GRADIENT IS POSITIVE
        if (m_V) {
            double err_now = calculateError(X); // Retrieve error
            if (_error < err_now) {
                System.out.println("broken out @" + e);
                break;
            }
            _error = err_now;
        }

        /*
         * The update
         */
        update(X);
    }

    return _error;
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Train - Setup and batch-train the RBM on X.
 * <br>//w  w w  . ja  v a 2 s . c  o  m
 * TODO, above function train(X_) could really be trained with train(X_,N), so, should share code with train(X)
 * <br>
 * TODO, divide gradient by the size of the batch! (doing already? .. no)
 * @param   X_         X
 * @param   batchSize   the batch size
 */
public double train(double X_[][], int batchSize) throws Exception {

    initWeights(X_);

    X_ = MatrixUtils.addBias(X_);

    int N = X_.length; // N
    if (batchSize == N)
        return train(X_);
    int N_n = (int) Math.ceil(N * 1. / batchSize);// Number of batches

    Matrix X_n[] = new Matrix[N_n];
    for (int n = 0, i = 0; n < N; n += batchSize, i++) {
        // @TODO, could save some small-time memory/speed here
        X_n[i] = new Matrix(Arrays.copyOfRange(X_, n, Math.min(n + batchSize, N)));
    }

    for (int e = 0; e < m_E; e++) {

        // @TODO could be random, see function below
        for (Matrix X : X_n) {
            update(X, 1. / N_n);
        }
    }

    return 1.0;
}

---

From source file:rbms.RBM.java

License:Open Source License

/**
 * Train - Setup and batch-train the RBM on X, with some random sampling involved.
 * <br>/*from   ww w.j  a  va 2s  .c  om*/
 * TODO should share code with train(X)
 * @param   X_         X
 * @param   batchSize   the batch size
 * @param   r         the randomness
 */
public double train(double X_[][], int batchSize, Random r) throws Exception {

    initWeights(X_);

    X_ = MatrixUtils.addBias(X_);
    int N = X_.length; // N
    int N_n = (int) Math.ceil(N * 1. / batchSize);// Number of batches

    // @TODO select the batches randomly at each epoch
    Matrix X_n[] = new Matrix[N_n];
    for (int n = 0, i = 0; n < N; n += batchSize, i++) {
        X_n[i] = new Matrix(Arrays.copyOfRange(X_, n, Math.min(n + batchSize, N)));
    }

    for (int e = 0; e < m_E; e++) {
        for (int i = 0; i < N_n; i++) {
            update(X_n[r.nextInt(N_n)]);
        }
    }

    return 1.0;
}

---