Adds a column vector to a transformed design matrix and updates its
QR-decomposition
Auxiliary function to compute quantiles of survival cdf
Complemented error function (double precision)
Complemented error function (single precision)
Determination of the parameter d of the Huber weight function
Initial values for the iterative algorithms implemented in
CYFALG, CYNALG, and CYGALG
Initial values for the Rocke estimates of covariance
Inverse of the cumulative Chi2-distribution function
Gives the current values ofthe parameters of the ROBETH
subroutine common blocks
Uniform random number generator
Exchanges two columns of a symmetric matrix
Diagonal matrix D for the Poisson case in discrete GLM
Auxiliary function for the computation of QQopt
Provide default values for most scalar parameters used by the Robeth subroutines
Sets default parameters for regression estimates
Fixed-point algorithm for the A-step
Determination of the parameter c of the Huber weight function from the
proportion eps of contamination
Determination of the correction factor for the M-estimate based on Huber
weight function
Resampling algorithm for the computation of S-estimates
Computation of fC.fC.inv(AT A) for a given matrix A and scale factor fC
Provide the current values of the scalar parameters used by the Robeth subroutines
Auxiliary function for the computation of QQopt
Set the current values of the scalar parameters used by the Robeth subroutines
Diagonal matrix D for the binomial case in discrete GLM
Cumulative F-distribution function
Logarithm at the Gamma-function at the point x
Newton-type algorithm for the c-step
Chi weight function for location and regression
Exchanges two columns of a symmetric matrix (double precision)
Parametric estimate of survival cdf
Resampling algorithm for the computation of weighted S-estimates
Assigns values to the ROBETH parameters included in common blocks
Forms the scalar (dot) product of two vectors
Newton-type algorithm for the computation of an
M-estimate of multivariate location and scatter
Binomial probability distribution
Initial values of theta, A and c_i,...,c_n
Asymptotic relative efficiency of a general M-estimate for a model
with mu quantitative and nu qualitative covariates
Fixed-point algorithm for the computation of an
M-estimate of multivariate location and scatter
Cumulative Gaussian distribution function (double precision)
Forms the scalar (dot) product of two vectors (double precision)
The total deviance of the fitted generalizrd linear model
Diagonal matrices D_G and E_G
Computation of Int Psi(s).Psi(s) dPhi(s) and Int Psi'(s) dPhi(s) when Psi is
a user-supplied external function
Computation of Beta0 = Phi_inv(0.75)
Conjugate gradient algorithm for the computation of an
M-estimate of multivariate location and scatter
Median and median absolute deviation
Main algorithm
Newton-type algorithm for the theta-step
M-estimate of location with simultaneous estimation of scale
Hodges-Lehman estimate and confidence intervals for the
center of symmetry based on the one-sample Wilcoxon test
Inverts a triangular matrix
Inverts the approximate null distribution of the
one-sample Wilcoxon test statistic
Inverts a triangular matrix (double precision)
Computation of Li, li and lip
Multiplies a lower-triangular matrix by its transpose (double precision)
Incomplete Gamma-integral function
Simultaneous computation of the MVE and LMS estimates
Multiplies a triangular matrix by a triangular matrix
Multiplies a full matrix by a full matrix
Diagonal matrices D_M, E_M, D_S, E_S when psi
is the Huber function
Interchanges two vectors (double precision)
Determination of the j-th order statistic
Diagonal matrices D_M, E_M, D_S, E_S when psi
is a user-supplied function
Logarithm of the Gamma-function at the point n/2
Backtransformation of the covariance matrix of the
coefficient estimates
Cumulative Gaussian distribution function
Evaluates a quadratic form (double precision)
S-algorithm for M-estimates
Multiplies a full matrix by a full matrix (double precision)
Classical estimates of mean and standard deviation
Computation of Int Psi(s).Psi(s) dPhi(s) and Int Psi'(s) dPhi(s) when Psi is
the Huber function
Cholesky decomposition of a symmetric matrix (double precision)
Print a message when a required argument is missing
Computes the diagonal elements of the hat matrix
Diagonal hat matrices D_M, E_M, D_S, and E_S
Computation of Int Chi(s) dPhi(s) when Chi=Psi.Psi/2 and Psi is
the Huber function
Multiplies a full matrix by a vector (double precision)
Computation of Int Chi(s) dPhi(s) when Chi is a user-supplied function
Constructs and/or applies a single elementary Householder transformation
(double precision)
Constructs and/or applies a single elementary Householder transformation
Median an median absolute deviation
Asymptotic relative efficiency of the tau-test
Forms the Euclidean norm of a vector
Computes the matrix Ktau
Resampling algorithm for the computation of the LMS estimate
Covariance matrix of the coefficient estimates of the
form f.inv(XT X)
Multiplies a symmetric matrix by a symmetric matrix (double precision)
Multiplies an upper-triangular matrix by its transpose
Resampling algorithm for the computation of the LTS estimate
u weight function for covariances
Algorithmic determination of the smallest double precision positive
number x
Diagonal 'check' matrices D_M, E_M, D_S, and E_S
Cumulative t-distribution function
Covariance matrix of the coefficient estimates of the
form f.inv(S1) S2 inv(S1)
Covariance matrix of the coefficient estimates of the form
f.inv(XT X) in the transformed coordinate system
Bounded influence regression
Correction factor f_H for the covariance matrix of a
Huber-type estimate
tau-test for the shift parameter
Nonparametric estimate and confidence intervals for the
shift parameter based on the Mann-Whitney test statistic
Forms the Euclidean norm of a vector (double precision)
Multiplies a symmetric matrix by a full matrix
Multiplies a lower-triangular matrix by a vector (double precision)
t-test for the shift parameter
Iterative algorithm for the computation of an M-estimate
of the scale parameter when the residuals are given
psi' weight function for location and regression
Inverts the approximate null distribution of the
Mann-Whitney test statistic
Multiplies a lower-triangular matrix by a vector
Cholesky decomposition of a symmetric matrix
Generation and comparison of all regressions on subsets
of covariates
w' weight function for covariances
Inverts the approximate null distribution of the sign test statistic
Dummy u user function
Solution of the least squares problem
Computation of (robust) t-statistics for t-directed search
W-algorithm for M-estimate of location
Multiplies a symmetric matrix by a full matrix (double precision)
Multiplies a full matrix by a vector
v weight function for covariances
u weight function for covariances
Poisson distribution
Multiplies a symmetric matrix by a full matrix when the result
is a symmetric matrix
Computation of the constant Beta0
Multiplies a symmetric matrix by a symmetric matrix
Solution of the least absolute residual problem
Multiplies an upper-triangular matrix by a vector
Multiplies a lower-triangular matrix by its transpose
Multiplies a symmetric matrix by a full matrix when the result
is a symmetric matrix
Cumulative distribution and density function of a linear
combination of chi-2 random variables
Zeng method for censored data
v' weight function for covariances
Computes the eigenvalues of the matrix Ktau
Algorithmic determination of the smallest double precision positive
number x
Computation of the constant Beta when Chi is a user-supplied function
Computation of the constant Beta when Chi=Psi.Psi/2 and Psi is the
Huber function
v weight function for covariances
Multiplies an upper-triangular matrix by its transpose (double precision)
H-algorithm for M-estimates
Interchanges two vectors
Multiplies a triangular matrix by a triangular matrix (double precision)
Convert local variable to Fortran character
Permutes the columns of the design matrix:
Predictors in omega are placed in the first q positions
Newton algorithm with adaptive steps for M-estimates
Multiplies an upper-triangular matrix by a vector
ub weight function for M-estimates in GLM
High breakdown point and high efficiency regression with test
for bias
Sorts the components of a vector in ascending order and permutes the
components of another vector accordingly
psi weight function for location and regression
psi weight function for location and regression
Fixed-point algorithm for the computation of the matrix A
w weight function for covariances
w weight function for covariances
u' weight function for covariances
Interface for the FORTRAN programs developed at the ETH-Zuerich and then at IUMSP-Lausanne
Permutes the elements of a vector
Permutes the columns of a matrix by means of transpositions
W-algorithm for M-estimates
Newton-Huber algorithm for the computation of the
lower-triangular matrix A in the standardized case
Evaluates a quadratic form
Removes a column from a transformed design matrix and
updates its QR-decomposition
psi' weight function for location and regression
Density of the norm of a standard Gaussian vector with
p components
Schweppe original weight proposal
Convert local variable to Fortran integer
Convert local variable to Fortran single precision
Scales a vector by a constant
Convert local variable to Fortran double precision
Inverse of the cumulative t-distribution function
Gaussian density function
Double precision version of RIMTRF
Upper triangularization (QR-decomposition) of the design
matrix and determination of its pseudorank
w weight function
Inverse of the standard Gaussian cumulative distribution function
Dummy u user function (double precision)
rho weight function for location and regression
Conjugate gradient algorithm for the computation of
the lower-triangular matrix A in the standardized case
Computes the Rn2-test statistic for a linear
hypothesis in canonical form
Sorts the components of a vector in ascending order
w' weight function for covariances
Rho weight function for location and regression
Scales a double precision vector by a constant
Computes the tau-test statistic for a linear
hypothesis in canonical form
Modified fixed-point algorithm for collinear
data in the standardized case
u' weight function for covariances
Initial value of the matrix A
Auxiliary function to find lambda parameter for extended loggamma distribution
Auxiliary function to find mu and sigma parameter for extended loggamma
distribution
Asymptotic relative efficiency of a general M-estimate for a model
with mu quantitative covariates with or without a constant term
Chi weight function for location and regression
Determination of the parameters a2 and b2 of the Huber weight function from the
proportion eps of contamination
Cumulative Chi-square distribution function