Example usage for org.apache.commons.math3.distribution TDistribution inverseCumulativeProbability

Introduction

In this page you can find the example usage for org.apache.commons.math3.distribution TDistribution inverseCumulativeProbability.

Prototype

public double inverseCumulativeProbability(final double p) throws OutOfRangeException 

Source Link

Document

The default implementation returns

• #getSupportLowerBound() for p = 0 ,
• #getSupportUpperBound() for p = 1 .

Usage

From source file:tinfour.gwr.SurfaceGwr.java

/**
 * Gets a value equal to one half of the range of the prediction interval
 * on the observed response at the interpolation coordinates for the
 * most recent call to computeRegression().
 *
 * @param alpha the significance level (typically 0..05, etc).
 * @return a positive value./*from w  ww .  j  a  v a  2 s. c  o  m*/
 */
public double getPredictionIntervalHalfRange(double alpha) {
    // TO DO: if the method is OLS, it would make sense to
    //        use a OLS version of this calculation rather than
    //        the more costly Leung version...  Also, I am not 100 %
    //        sure that they converge to the same answer, though they should
    computeVarianceAndHat();
    //double effDegOfF = getEffectiveDegreesOfFreedom(); // should match delta1

    double[][] input = computeDesignMatrix(model, xOffset, yOffset, nSamples, samples);
    RealMatrix mX = new BlockRealMatrix(input);
    RealMatrix mXT = mX.transpose();

    // the weights array may not necessarily be of dimension nSamples,
    // so we need to copy it
    double[] rW = Arrays.copyOf(weights, nSamples);
    RealMatrix mW = new DiagonalMatrix(rW);
    RealMatrix design = mXT.multiply(mW).multiply(mX);
    RealMatrix vcm;
    try {
        QRDecomposition cd = new QRDecomposition(design);
        DecompositionSolver s = cd.getSolver();
        vcm = s.getInverse();
    } catch (SingularMatrixException npex) {
        return Double.NaN;
    }

    double nLeungDOF = (delta1 * delta1 / delta2);

    for (int i = 0; i < nSamples; i++) {
        rW[i] = weights[i] * weights[i];
    }

    DiagonalMatrix mW2 = new DiagonalMatrix(rW);
    RealMatrix mS = vcm.multiply(mXT).multiply(mW2).multiply(mX).multiply(vcm);
    double pS = mS.getEntry(0, 0);
    double p = Math.sqrt(this.sigma2 * (1 + pS));

    TDistribution td = new TDistribution(nLeungDOF);

    double ta = td.inverseCumulativeProbability(1.0 - alpha / 2.0);

    return ta * p;
}

---

From source file:tools.descartes.bungee.evaluation.ScalabilityReproducibilityEvaluation.java

private double getConfidenceIntervalWidth(SummaryStatistics summaryStatistics, double confidence) {
    double significance = 1 - confidence;
    TDistribution tDist = new TDistribution(summaryStatistics.getN() - 1);
    double a = tDist.inverseCumulativeProbability(1.0 - significance / 2);
    return a * summaryStatistics.getStandardDeviation() / Math.sqrt(summaryStatistics.getN());
}

---

From source file:UserPerformance.GetLowerBound.java

private static double calcMeanCI(double n, double StandardDeviation, double alfa) {
    try {/*from ww  w  . j  a va 2s  . c om*/
        // Create T Distribution with N-1 degrees of freedom
        TDistribution tDist = new TDistribution(n - 1);
        // Calculate critical value
        double critVal = tDist.inverseCumulativeProbability(1.0 - (1 - alfa) / 2);
        // Calculate confidence interval
        return (critVal * StandardDeviation / Math.sqrt(n));
    } catch (MathIllegalArgumentException e) {
        return Double.NaN;
    }
}

---

From source file:wsattacker.plugin.intelligentdos.StatisticTest.java

private static double calcMeanCI(SummaryStatistics stats, double level) {
    // Create T Distribution with N-1 degrees of freedom
    TDistribution tDist = new TDistribution(stats.getN() - 1);
    // Calculate critical value
    double critVal = tDist.inverseCumulativeProbability(1.0 - (1 - level) / 2);
    // Calculate confidence interval
    return critVal * stats.getStandardDeviation() / Math.sqrt(stats.getN());
}

---

• «
• 1
• 2
• 3
• »