An example of a general implementation of a parameter optimization scheme using core features of the copBasic package. Because of the general complexity of the objectives for this function, the interface shown here is considered an “beta version” and nomenclature is subject to possibly sweeping changes in the future.
copBasic.fitpara.beta(uv=NULL, popstat=NULL, statf=NULL, cop=NULL,
paradim=1, interval=NULL, par.init=NULL, ...)A vector of the values for the parameter variable is returned
An R two column matrix or data.frame of a sample of nonexceedance probabilities \(u\) and \(v\);
The population statistic(s) that will be used if uv is NULL;
A function responsible at the minimum for computation of the theoretical values of the population statistic(s) that the optimization will revolve around; This function, if supporting an as.sample interface (e.g. hoefCOP) and if uv has been provided, will be dispatched to compute the population statistic(s);
A copula function that is passed along to statf though of course the statf function can decide whether to use this argument or not;
The dimension of the parameters. In reality, the default triggers uni-dimensional root solving using the uniroot() function in R or otherwise the optim() function in R is used for multi-dimensional minimization with subtle changes in setup (see source code). Alternative logic could be have been used but it is felt that this is the most logical for future adaptations;
The interval argument by the same name for the uniroot() function;
The initial parameter vector for the optim() function; and
Additional arguments to pass.
W.H. Asquith