Class ProbabilityCorrelation

java.lang.Object
com.polytechnik.utils.ProbabilityCorrelation

public class ProbabilityCorrelation extends Object
Probability(or Value) Correlation of p and i. Actually is the case d=2 of LebesgueQuadratureWithEVData, the corr is the difference of diagonal and off-diagonal elemets: (diag-offdiag)/(diag+offdiag) elemets, depnding of flag_probabiliy_correlation flag, of the matrix returned by either getProbabilityCorrelation or getValueCorrelation of LebesgueQuadratureWithEVData.
  • Field Details

    • pL

      public final double pL
      Min/Max eigenvalue.
    • pH

      public final double pH
      Min/Max eigenvalue.
    • iL

      public final double iL
      Min/Max eigenvalue.
    • iH

      public final double iH
      Min/Max eigenvalue.
    • W

      public final double W
      W=w_pL_iL+w_pL_iH+w_pH_iL+w_pH_iH. corr=(w_pL_iL-w_pL_iH-w_pH_iL+w_pH_iH)/W.
    • w_pL_iL

      public final double w_pL_iL
      W=w_pL_iL+w_pL_iH+w_pH_iL+w_pH_iH. corr=(w_pL_iL-w_pL_iH-w_pH_iL+w_pH_iH)/W.
    • w_pL_iH

      public final double w_pL_iH
      W=w_pL_iL+w_pL_iH+w_pH_iL+w_pH_iH. corr=(w_pL_iL-w_pL_iH-w_pH_iL+w_pH_iH)/W.
    • w_pH_iL

      public final double w_pH_iL
      W=w_pL_iL+w_pL_iH+w_pH_iL+w_pH_iH. corr=(w_pL_iL-w_pL_iH-w_pH_iL+w_pH_iH)/W.
    • w_pH_iH

      public final double w_pH_iH
      W=w_pL_iL+w_pL_iH+w_pH_iL+w_pH_iH. corr=(w_pL_iL-w_pL_iH-w_pH_iL+w_pH_iH)/W.
    • corr

      public final double corr
      W=w_pL_iL+w_pL_iH+w_pH_iL+w_pH_iH. corr=(w_pL_iL-w_pL_iH-w_pH_iL+w_pH_iH)/W.
    • flag_probabiliy_correlation

      public final boolean flag_probabiliy_correlation
      If true, probability correlation is calculated, otherwise value correlation.
  • Constructor Details

    • ProbabilityCorrelation

      public ProbabilityCorrelation(double[] QQ, double[] QQp, double[] QQi)
      Obtain probability correlation from three 2x2 matrices: https://arxiv.org/abs/1709.06759, Appendix C: "Probability Correlation of Variables". For probability correlation (flag_probabiliy_correlation=true), no basis functions required, all we need is eigenvalues problem solver M.EV, use a standard one. This is not the case for value correlation.
    • ProbabilityCorrelation

      public ProbabilityCorrelation(double x0, double x1, double x2, double px0, double px1, double px2, double ix0, double ix1, double ix2, boolean flag_probabiliy_correlation)
      Probaility Correlation (from monomials moments).
    • ProbabilityCorrelation

      public ProbabilityCorrelation(double x0, double x1, double x2, double px0, double px1, double px2, double ix0, double ix1, double ix2, OrthogonalPolynomialsBasisFunctionsCalculatable<? extends BasisFunctionsCalculatable> M, boolean flag_probabiliy_correlation)
      Probaility Correlation (from the moments in arbitrary basis, the Q_0(x) is assumed to be a constant!).
    • ProbabilityCorrelation

      protected ProbabilityCorrelation(double x0, double x1, double x2, double px0, double px1, double px2, double ix0, double ix1, double ix2, OrthogonalPolynomialsBasisFunctionsCalculatable<? extends BasisFunctionsCalculatable> M, boolean flag_probabiliy_correlation, double eps_if_selftest_required)
      Probaility Correlation (from the moments in arbitrary basis, the Q_0(x) is assumed to be a constant!).
    • ProbabilityCorrelation

      public ProbabilityCorrelation(double[] QQ, double[] QQp, double[] QQi, OrthogonalPolynomialsBasisFunctionsCalculatable<? extends BasisFunctionsCalculatable> M, boolean flag_probabiliy_correlation)
      From 2x2 matrices in arbitraty basis, the Q_0(x) is assumed to be a constant, otherwise averages will be incorrect. Unit selftest is disabled.
    • ProbabilityCorrelation

      protected ProbabilityCorrelation(double[] QQ, double[] QQp, double[] QQi, OrthogonalPolynomialsBasisFunctionsCalculatable<? extends BasisFunctionsCalculatable> M, boolean flag_probabiliy_correlation, double eps_if_selftest_required)
      From 2x2 matrices in arbitraty basis, if eps_if_selftest_required>0, then unit self-test with this eps is run.
  • Method Details