List of usage examples for weka.core.matrix Matrix Matrix
public Matrix(double[][] A, int m, int n)
From source file:mulan.classifier.neural.ThresholdFunction.java
License:Open Source License
/** * Build a threshold function for based on input data. * The threshold function is build for a particular model. * * @param idealLabels the ideal output for each input patterns, which a model should output. * First index is expected to be number of examples and second is the label index. * @param modelOutLabels the real output of a model for each input pattern. * First index is expected to be number of examples and second is the label index. * @throws IllegalArgumentException if dimensions of input arrays does not match *///w ww .jav a 2 s . c o m public void build(final double[][] idealLabels, final double[][] modelOutLabels) { if (idealLabels == null || modelOutLabels == null) { throw new IllegalArgumentException("Non of the input parameters can be null."); } int numExamples = idealLabels.length; int numLabels = idealLabels[0].length; if (modelOutLabels.length != numExamples || modelOutLabels[0].length != numLabels) { throw new IllegalArgumentException("Matrix dimensions of input parameters does not agree."); } double[] thresholds = new double[numExamples]; double[] isLabelModelOuts = new double[numLabels]; double[] isNotLabelModelOuts = new double[numLabels]; for (int example = 0; example < numExamples; example++) { Arrays.fill(isLabelModelOuts, Double.MAX_VALUE); Arrays.fill(isNotLabelModelOuts, -Double.MAX_VALUE); for (int label = 0; label < numLabels; label++) { if (idealLabels[example][label] == 1) { isLabelModelOuts[label] = modelOutLabels[example][label]; } else { isNotLabelModelOuts[label] = modelOutLabels[example][label]; } } double isLabelMin = isLabelModelOuts[Utils.minIndex(isLabelModelOuts)]; double isNotLabelMax = isNotLabelModelOuts[Utils.maxIndex(isNotLabelModelOuts)]; // check if we have unique minimum ... // if not take center of the segment ... if it is a segment if (isLabelMin != isNotLabelMax) { // check marginal cases -> all labels are in or none of them if (isLabelMin == Double.MAX_VALUE) { thresholds[example] = isNotLabelMax + 0.1; } else if (isNotLabelMax == -Double.MAX_VALUE) { thresholds[example] = isLabelMin - 0.1; } else { // center of a segment thresholds[example] = (isLabelMin + isNotLabelMax) / 2; } } else { // when minimum is unique thresholds[example] = isLabelMin; } } Matrix modelMatrix = new Matrix(numExamples, numLabels + 1, 1.0); modelMatrix.setMatrix(0, numExamples - 1, 0, numLabels - 1, new Matrix(modelOutLabels)); Matrix weights = modelMatrix.solve(new Matrix(thresholds, thresholds.length)); double[][] weightsArray = weights.transpose().getArray(); parameters = Arrays.copyOf(weightsArray[0], weightsArray[0].length); }
From source file:org.mitre.clustering.AffinityPropagation.java
License:Open Source License
public AffinityPropagation(Matrix sims, int max, int cons, double lambda, double p) { s = sims;/*from ww w.j ava 2 s . c o m*/ maxits = max; convits = cons; lam = lambda; count = s.getColumnDimension(); /* * Add noise to get rid of degeneracies */ Matrix rand = Matrix.random(count, count); Matrix small = new Matrix(count, count, .0000001); Matrix t = s.times(Double.MIN_VALUE).plus(small); t = t.times(rand); s = s.plus(t); /* * Put preferences on diagonal of S */ for (int i = 0; i < count; i++) { s.set(i, i, p); } try { run(); } catch (Exception ex) { ex.printStackTrace(); } }
From source file:org.mitre.clustering.AffinityPropagation.java
License:Open Source License
public void run() throws Exception { /*/* w w w . j ava2 s .c om*/ * Allocate space for messages */ Matrix ds; try { ds = getDiag(s); } catch (Exception ex) { ex.printStackTrace(); } A = new Matrix(count, count, 0.0); R = new Matrix(count, count, 0.0); int t = 1; /* * Run parralel affinity prop. updates */ Matrix e = new Matrix(count, convits, 0); Boolean dn = false; int i = -1; // adjusted due to difference in MATLAB indexing Matrix st; if (!symmetric) { st = s; } else { st = s.transpose(); } while (!dn) { i = i + 1; A = A.transpose(); R = R.transpose(); /* * Compute responsibilities */ for (int ii = 0; ii < count; ii++) { Matrix old = R.getMatrix(0, count - 1, ii, ii); Matrix AS = A.getMatrix(0, count - 1, ii, ii).plus(st.getMatrix(0, count - 1, ii, ii)); int I = getMaxIndex(AS); double Y = AS.get(I, 0); /* * Need to use -Double.MAX_VALUE here, as this will return * the smallest NEGATIVE number, as opposed to smallest absolute * floating point number */ AS.set(I, 0, -Double.MAX_VALUE); int I2 = getMaxIndex(AS); double Y2 = AS.get(I2, 0); /* * Set R */ Matrix yMat = new Matrix(count, 1, Y); Matrix stVect = st.getMatrix(0, count - 1, ii, ii); R.setMatrix(0, count - 1, ii, ii, stVect.minus(yMat)); R.set(I, ii, st.get(I, ii) - Y2); Matrix rVect = R.getMatrix(0, count - 1, ii, ii); rVect = rVect.times(1 - lam).plus(old.times(lam)); R.setMatrix(0, count - 1, ii, ii, rVect); } A = A.transpose(); R = R.transpose(); /* * Compute availabilities */ for (int jj = 0; jj < count; jj++) { Matrix old = A.getMatrix(0, count - 1, jj, jj); Matrix Rp = new Matrix(count, 1, 0); for (int c = 0; c < count; c++) { if (R.get(c, jj) < 0.0) { Rp.set(c, 0, 0.0); } else { Rp.set(c, 0, R.get(c, jj)); } } Rp.set(jj, 0, R.get(jj, jj)); Matrix sumM = new Matrix(count, 1, getSum(Rp)); A.setMatrix(0, count - 1, jj, jj, sumM.minus(Rp)); double dA = A.get(jj, jj); Matrix aVect = A.getMatrix(0, count - 1, jj, jj); for (int c = 0; c < count; c++) { if (aVect.get(c, 0) > 0) { aVect.set(c, 0, 0.0); } } A.setMatrix(0, count - 1, jj, jj, aVect); A.set(jj, jj, dA); aVect = A.getMatrix(0, count - 1, jj, jj); aVect = aVect.times(1 - lam).plus(old.times(lam)); A.setMatrix(0, count - 1, jj, jj, aVect); } /* * Check for convergence */ Matrix E = null; try { E = getDiag(A).plus(getDiag(R)); } catch (Exception ex) { ex.printStackTrace(); } for (int c = 0; c < count; c++) { if (E.get(c, 0) > 0) { E.set(c, 0, 1.0); } else { E.set(c, 0, 0.0); } } int index = i % convits; e.setMatrix(0, count - 1, index, index, E); K = getSum(E); if (i >= convits || i >= maxits) { Matrix se = new Matrix(count, 1); for (int c = 0; c < count; c++) { double sum = 0; for (int c2 = 0; c2 < convits; c2++) { sum += e.get(c, c2); } se.set(c, 0, sum); } Matrix seC = matrixIsEqual(se, convits); Matrix seZ = matrixIsEqual(se, 0.0); double value = getSum(seC.plus(seZ)); if (value == count) { unconverged = false; } else { unconverged = true; } if ((!unconverged && K > 0) || i == maxits) { dn = true; } } } /* * Identify exemplars */ Matrix diag = null; try { diag = getDiag(A).plus(getDiag(R)); } catch (Exception ex) { ex.printStackTrace(); } exemplars = new Vector<Integer>(); for (int c = 0; c < count; c++) { if (diag.get(c, 0) > 0) { exemplars.add(c); } } }
From source file:org.mitre.clustering.AffinityPropagation.java
License:Open Source License
/** * Returns the diagonal of a square matrix * //w w w . j a v a 2s .co m * @param m * @return */ private Matrix getDiag(Matrix m) throws Exception { if (m.getRowDimension() != m.getColumnDimension()) { throw new Exception("Matrix must be square"); } int c = m.getRowDimension(); Matrix rm = new Matrix(c, 1, 0); for (int i = 0; i < c; i++) { rm.set(i, 0, m.get(i, i)); } return rm; }