R: Auxiliary for Controlling Qest Fitting

Qest.control {Qest}

R Documentation

Auxiliary for Controlling Qest Fitting

Description

Auxiliary function for controlling Qest fitting. Estimation proceeds in three steps: (i) evaluation of starting points; (iia) stochastic gradient-based optimization (iib) standard gradient-based optimization; and (iii) Newton-Raphson. Step (i) is initialized at the provided starting values (the start argument of Qest), and utilizes a preliminary flexible model, estimated with pchreg, to generate a cheap guess of the model parameters. If you have good starting points, you can skip step (i) by setting restart = FALSE. Steps (iia) and (iib) find an approximate solution, and make sure that the Jacobian matrix is well-defined. Finally, step (iii) finds a more precise solution.

Usage

Qest.control(tol = 1e-8, maxit, safeit, alpha0, display = FALSE, restart = FALSE)

Arguments

`tol`	tolerance for convergence of Newton-Raphson algorithm, default is 1e-8.
`maxit`	maximum number of iterations of Newton-Raphson algorithm. If not provided, a default is computed as `50 + 25*npar`, where `npar` is the number of parameters.
`safeit`	maximum number of iterations of gradient-search algorithm. If not provided, a default is computed as `10 + 5*npar`, where `npar` is the number of parameters.
`alpha0`	step size for the preliminary gradient-based iterations. If estimation fails, you can try choosing a small value of `alpha0`. If `alpha0` is missing, an adaptive choiche will be made internally.
`display`	Logical. If `TRUE`, tracing information on the progress of the optimization is printed on screen. Default is `FALSE`.
`restart`	Logical. If `FALSE` (the default), step (i) is not performed, and the provided starting points are directly passed to step (ii). This may save you some time, but is not recommended unless you are confident about your choice of initial values. When `restart = TRUE`, the provided starting points are used to initialize step (i).

Details

If called with no arguments, Qest.control() returns a list with the current settings of these parameters. Any arguments included in the call sets those parameters to the new values, and then silently returns.

Value

A list with named elements as in the argument list

Note

Step (i) is not performed, and restart is ignored, if the quantile function is one of the available Qfamily.

Author(s)

Gianluca Sottile <gianluca.sottile@unipa.it> Paolo Frumento <paolo.frumento@unipi.it>