org.ujmp.core.Matrix.toColumnVector()方法的使用及代码示例

x33g5p2x  于2022-01-25 转载在 其他  
字(8.6k)|赞(0)|评价(0)|浏览(93)

本文整理了Java中org.ujmp.core.Matrix.toColumnVector()方法的一些代码示例，展示了Matrix.toColumnVector()的具体用法。这些代码示例主要来源于Github/Stackoverflow/Maven等平台，是从一些精选项目中提取出来的代码，具有较强的参考意义，能在一定程度帮忙到你。Matrix.toColumnVector()方法的具体详情如下：
包路径：org.ujmp.core.Matrix
类名称：Matrix
方法名：toColumnVector

Matrix.toColumnVector介绍

暂无

代码示例

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
  input = input.toColumnVector(Ret.NEW);
  Matrix input2 = Matrix.Factory.zeros(1, selectedFeatures.size());
  for (int i = 0; i < selectedFeatures.size(); i++) {
    input2.setAsDouble(input.getAsDouble(0, selectedFeatures.get(i)), 0, i);
  }
  return learningAlgorithm.predictOne(input2);
}

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix hash(Matrix input) {
  Matrix result = input.toColumnVector(Ret.NEW).mtimes(randomVectors).ge(Ret.NEW, 0);
  return result;
}

代码示例来源：origin: jdmp/java-data-mining-package

public final int getClassCount() {
  return (int) get(0).getAsMatrix(TARGET).toColumnVector(Ret.LINK).getColumnCount();
}

代码示例来源：origin: jdmp/java-data-mining-package

public final int getFeatureCount() {
  return (int) get(0).getAsMatrix(INPUT).toColumnVector(Ret.LINK).getColumnCount();
}

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix compress(Matrix input) {
  return input.toColumnVector(Ret.LINK).minus(Ret.LINK, true, mean).divide(std).mtimes(u);
}

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
  input = input.toColumnVector(Ret.LINK);
  final double[] probs = new double[classCount];
  final double[] logs = new double[classCount];
  for (int j = 0; j < classCount; j++) {
    logs[j] += Math.log(classDists[j].getProbability(1.0));
  }
  // for all features
  for (int j = 0; j < input.getColumnCount(); j++) {
    // for all classes
    double probSum = 0;
    for (int i = 0; i < classCount; i++) {
      double value = input.getAsDouble(0, j);
      double probability = dists[j][i].getProbability(value);
      probs[i] = probability;
      probSum += probability;
    }
    for (int i = 0; i < classCount; i++) {
      logs[i] += Math.log(probs[i] / probSum);
    }
  }
  final double[] finalProbs = MathUtil.logToProbs(logs);
  Matrix m = Matrix.Factory.linkToArray(finalProbs).transpose();
  return m;
}

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
    List<Sortable<Double, Matrix>> bestResults = new FastArrayList<Sortable<Double, Matrix>>();
    for (Sample s : dataSet) {
      Matrix reference = s.getAsMatrix(getInputLabel());
      double distance = input.euklideanDistanceTo(reference, true);
      if (bestResults.size() < k) {
        bestResults.add(new Sortable<Double, Matrix>(distance, s
            .getAsMatrix(getTargetLabel())));
        Collections.sort(bestResults);
      } else if (distance < bestResults.get(k - 1).getComparable()) {
        bestResults.remove(k - 1);
        bestResults.add(new Sortable<Double, Matrix>(distance, s
            .getAsMatrix(getTargetLabel())));
        Collections.sort(bestResults);
      }
    }
    List<Matrix> results = new FastArrayList<Matrix>();
    for (Sortable<Double, Matrix> s : bestResults) {
      results.add(s.getObject().toColumnVector(Ret.LINK));
    }
    Matrix resultMatrix = Matrix.Factory.vertCat(results);
    Matrix mean = resultMatrix.mean(Ret.NEW, Matrix.ROW, true);
    return mean;
  }
}

代码示例来源：origin: jdmp/java-data-mining-package

System.out.println("Converting samples: " + Math.round(((double) i / dataSet.size() * 100)) + "% done");
Matrix input = s.getAsMatrix(getInputLabel()).toColumnVector(Ret.LINK);
int targetClass = (int) s.getAsMatrix(getTargetLabel()).toColumnVector(Ret.LINK).getCoordinatesOfMaximum()[COLUMN];
prob.y[i] = targetClass;
long columnCount = input.getColumnCount();

代码示例来源：origin: jdmp/java-data-mining-package

public final void predictOne(Sample sample) {
  Matrix predicted = predictOne(sample.getAsMatrix(getInputLabel()));
  predicted = predicted.toColumnVector(Ret.LINK);
  sample.put(PREDICTED, predicted);
  if (sample.getAsMatrix(getTargetLabel()) != null) {
    Matrix target = sample.getAsMatrix(getTargetLabel()).toColumnVector(Ret.LINK);
    Matrix error = target.minus(predicted);
    sample.put(DIFFERENCE, error);
    sample.put(RMSE, Matrix.Factory.linkToValue(error.getRMS()));
  }
  sample.fireValueChanged();
}

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
  input = input.toColumnVector(Ret.NEW);
  Matrix bias = Matrix.Factory.ones(input.getRowCount(), 1);
  Matrix data = Matrix.Factory.horCat(bias, input);
  Matrix result = getParameterMatrix().transpose().mtimes(data.transpose());
  return result.transpose();
}

代码示例来源：origin: jdmp/java-data-mining-package

public void trainAll(ListDataSet dataSet) {
  selectFeatures(dataSet);
  ListDataSet newDataSet = new DefaultListDataSet();
  for (Sample s1 : dataSet) {
    Sample s2 = new DefaultSample();
    Matrix input1 = s1.getAsMatrix(getInputLabel()).toColumnVector(Ret.NEW);
    Matrix input2 = Matrix.Factory.zeros(1, selectedFeatures.size());
    for (int i = 0; i < selectedFeatures.size(); i++) {
      input2.setAsDouble(input1.getAsDouble(0, selectedFeatures.get(i)), 0, i);
    }
    s2.put(getInputLabel(), input2);
    s2.put(getTargetLabel(), s1.getAsMatrix(getTargetLabel()));
    newDataSet.add(s2);
  }
  learningAlgorithm.trainAll(newDataSet);
}

代码示例来源：origin: jdmp/java-data-mining-package

private Matrix createCompleteMatrix(ListDataSet dataSet) {
  final int sampleCount = dataSet.size();
  final int featureCount = getFeatureCount(dataSet);
  final int targetCount = getClassCount(dataSet);
  Matrix m = Matrix.Factory.zeros(sampleCount, featureCount + targetCount);
  for (int r = 0; r < sampleCount; r++) {
    Sample s = dataSet.get(r);
    Matrix input = s.getAsMatrix(getInputLabel()).toColumnVector(Ret.NEW);
    Matrix target = s.getAsMatrix(getTargetLabel()).toColumnVector(Ret.NEW);
    for (int c = 0; c < featureCount; c++) {
      m.setAsDouble(input.getAsDouble(0, c), r, c);
    }
    for (int c = 0; c < targetCount; c++) {
      m.setAsDouble(target.getAsDouble(0, c), r, c + featureCount);
    }
  }
  return m;
}

代码示例来源：origin: jdmp/java-data-mining-package

for (Sample s : dataSet) {
  final Matrix sampleInput = s.getAsMatrix(getInputLabel()).toColumnVector(Ret.LINK);
  final Matrix sampleTarget = s.getAsMatrix(getTargetLabel()).toColumnVector(Ret.LINK);
  final double weight = s.getWeight();

代码示例来源：origin: jdmp/java-data-mining-package

public void trainAll(ListDataSet dataSet) {
  featureCount = getFeatureCount(dataSet);
  classCount = getClassCount(dataSet);
  dimensions = featureCount + classCount;
  Matrix x = Matrix.Factory.zeros(dataSet.size(), dimensions);
  int i = 0;
  for (Sample s : dataSet) {
    Matrix input = s.getAsMatrix(getInputLabel()).toColumnVector(Ret.LINK);
    for (int c = 0; c < featureCount; c++) {
      x.setAsDouble(input.getAsDouble(0, c), i, c);
    }
    Matrix target = s.getAsMatrix(getTargetLabel()).toColumnVector(Ret.LINK);
    for (int c = 0; c < classCount; c++) {
      x.setAsDouble(target.getAsDouble(0, c), i, c + featureCount);
    }
    i++;
  }
  meanMatrix = x.mean(Ret.NEW, Matrix.ROW, true);
  covarianceMatrix = x.cov(Ret.NEW, true, true);
  try {
    inverse = covarianceMatrix.inv();
    factor = 1.0 / Math.sqrt(covarianceMatrix.det() * Math.pow(2.0 * Math.PI, dimensions));
  } catch (Exception e) {
    inverse = covarianceMatrix.pinv();
    factor = 1.0;
  }
}

代码示例来源：origin: jdmp/java-data-mining-package

double sampleWeight = MathUtil.getMatrix(matrices.get(SAMPLEWEIGHT)).doubleValue();
Matrix transposedInput = MathUtil.getMatrix(matrices.get(INPUT)).toColumnVector(Ret.NEW);

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
  input = input.toColumnVector(Ret.LINK);
  long columnCount = input.getColumnCount();
  int count = 0;
  for (int j = 0; j < columnCount; j++) {
    double value = input.getAsDouble(0, j);
    if (value != 0.0 && !MathUtil.isNaNOrInfinite(value)) {
      count++;
    }
  }
  Feature[] x = new Feature[count];
  count = 0;
  for (int j = 0; j < columnCount; j++) {
    double value = input.getAsDouble(0, j);
    if (value != 0.0 && !MathUtil.isNaNOrInfinite(value)) {
      x[count] = features.get(j + 1, value);
      count++;
    }
  }
  return predictOne(x);
}

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
  input = input.toColumnVector(Ret.NEW);
  Matrix x = Matrix.Factory.zeros(1, input.getColumnCount() + 1);
  for (int c = 0; c < input.getColumnCount(); c++) {
    x.setAsDouble(input.getAsDouble(0, c), 0, c + 1);
  }
  x = x.minus(mean).divide(std);
  x.setAsDouble(1, 0, 0);
  Matrix result = x.mtimes(getParameterMatrix());
  return result;
}

代码示例来源：origin: jdmp/java-data-mining-package

int randomIndex = MathUtil.nextInteger(dataSet.size());
Sample s = dataSet.get(randomIndex);
inputs.add(s.getAsMatrix(Sample.INPUT).toColumnVector(Ret.NEW));
targets.add(s.getAsMatrix(Sample.TARGET).toColumnVector(Ret.NEW));

代码示例来源：origin: jdmp/java-data-mining-package

public Matrix predictOne(Matrix input) {
  input = input.toColumnVector(Ret.NEW);
  Matrix x = Matrix.Factory.zeros(1, dimensions);
  for (int i = 0; i < featureCount; i++) {
    x.setAsDouble(input.getAsDouble(0, i), 0, i);
  }
  Matrix result = Matrix.Factory.zeros(1, classCount);
  double sum = 0;
  for (int i = 0; i < classCount; i++) {
    if (i > 0) {
      x.setAsDouble(0, 0, featureCount + i - 1);
    }
    x.setAsDouble(1, 0, featureCount + i);
    double density = getDensity(x);
    result.setAsDouble(density, 0, i);
    sum += density;
  }
  for (int i = 0; i < classCount; i++) {
    result.setAsDouble(result.getAsDouble(0, i) / sum, 0, i);
  }
  return result;
}

代码示例来源：origin: jdmp/java-data-mining-package

Matrix input = s.getAsMatrix(getInputLabel()).toColumnVector(Ret.LINK);
for (int c = 0; c < input.getColumnCount(); c++) {
  x.setAsDouble(input.getAsDouble(0, c), i, c);

内容来源于网络，如有侵权，请联系作者删除！

Matrix

相关文章

    查看更多

    热门标签

    更多
    JavaquerypythonNode开发语言requestUtil数据库Table后端算法LoggerMessageElementParser

    最新文章

    更多

    Matrix类方法

    • 蜀ICP备2022004421号-2 NULL123 https://www.null123.com © All rights reserved
    • 本站内容来源互联网,如果侵犯您的权益请联系我们删除, 联系方式：448109455@qq.com