fstats_robust_statistics Module



Interfaces

public interface median_absolute_deviation

  • private pure function median_absolute_deviation_array(x) result(rst)

    Computes the median absolute deviation of a data set by determining the median of distances of each data point from the data set median.

    Arguments

    Type IntentOptional Attributes Name
    real(kind=real64), intent(in), dimension(:) :: x

    The data set.

    Return Value real(kind=real64)

    The median absolute deviation of x.

  • private pure function median_absolute_deviation_column(x) result(rst)

    Computes median absolute deviation using the array implementation on a copy. Distances that overflow are saturated as in the array overload.

    Arguments

    Type IntentOptional Attributes Name
    class(data_column), intent(in) :: x

    The input column, unchanged; unallocated storage is treated as empty.

    Return Value real(kind=real64)

    Unscaled median absolute deviation; empty/nonfinite data give NaN.


Functions

public pure elemental function tukey_biweight_irls_weight(u, k) result(rst)

Computes Tukey's biweight IRLS weight for a standardized residual. The positive tuning constant k defaults to 4.685.

Arguments

Type IntentOptional Attributes Name
real(kind=real64), intent(in) :: u

The standardized residual.

real(kind=real64), intent(in), optional :: k

The positive cutoff/tuning constant; defaults to 4.685.

Return Value real(kind=real64)

The nonnegative IRLS weight.

public pure elemental function tukey_biweight_psi(u, k) result(rst)

Computes Tukey's biweight score function for a standardized residual. The positive tuning constant k defaults to 4.685.

Arguments

Type IntentOptional Attributes Name
real(kind=real64), intent(in) :: u

The standardized residual.

real(kind=real64), intent(in), optional :: k

The positive cutoff/tuning constant; defaults to 4.685.

Return Value real(kind=real64)

The score value psi(u).

public pure elemental function tukey_biweight_rho(u, k) result(rst)

Computes Tukey's biweight rho loss for a standardized residual. The positive tuning constant k defaults to 4.685.

Arguments

Type IntentOptional Attributes Name
real(kind=real64), intent(in) :: u

The standardized residual.

real(kind=real64), intent(in), optional :: k

The positive cutoff/tuning constant; defaults to 4.685.

Return Value real(kind=real64)

The loss value rho(u).