Control for the Robust Gamma Parameter Estimator

DESCRIPTION:

Generates a list of control parameters for the robust gamma distribution parameter estimator used in gammaRob.

USAGE:

gammaRob.control(estim, b1=1.5, b2=1.7, k=101, A=c(0, 0, 0),
                 maxta=1, maxtc=1, maxit=100, tol=0.0001, til=0.001,
                 sigma=0, alpha1=0.5, alpha2=20.5, u=0.99, beta=0.4,
                 gam=0.4, cov=T)

REQUIRED ARGUMENTS:

estim
the robust estimator used in gammaRob. The choices are "tdmean" for the truncated mean and "M" for the M. The M estimator is not available for the lognormal distribution. Only the parameters required for this estimator are returned.

Parameters for the truncated mean:

alpha1
the regula falsi procedure (Marazzi and Ruffieux 1999) is applied to the interval [alpha1, alpha2].

alpha2
see alpha1.

u
the value of the tuning constant.

beta
value of Beta in the interval [0, 0.5].

gam
value of Gamma in the interval [0, 0.5].

tol
relative precision of alpha, sigma, and the truncation points.

cov
if TRUE an estimate of the variance of the mean is computed.

Parameters for the M:

b1
the tuning constant b1.

b2
the tuning constant b2.

k
the number of values of alpha for which Ab(alpha) and cb(alpha) must be computed.

A
initial value of Ab(alpha).

maxta
maximum number of steps for improving Ab(alpha).

maxtc
maximum number of steps for improving cb(alpha).

maxit
maximum number of cycles of the main algorithm.

tol
desired relative precision of each element of Ab(alpha) and cb(alpha).

til
desired relative accuracy of numerical integrations.

sigma
if positive, then the only parameter fit is alpha.

cov
if TRUE the covariance matrix of the parameters is returned.

VALUE:

a list containing the control parameters for the estimator specified by estim .

VALUE:

SEE ALSO:

.

EXAMPLES:

gammaRob(los, control = gammaRob.control("tdmean", tol=1e-6))