R: Optimizes the Composite log-likelihood

CompLik {GeoModels}

R Documentation

Optimizes the Composite log-likelihood

Description

Subroutine called by GeoFit. The procedure estimates the model parameters by maximisation of the composite log-likelihood.

Usage

CompLik(copula,bivariate, coordx, coordy ,coordt, 
coordx_dyn, corrmodel, data,distance,  flagcorr, 
flagnuis, fixed, GPU, grid,likelihood,local,lower, 
                 model, n, namescorr, namesnuis, 
                 namesparam,
               numparam, numparamcorr, optimizer, 
               onlyvar, parallel, param,
               spacetime, type,upper, varest, vartype,
               weigthed, winconst, winstp,winconst_t, 
               winstp_t, ns, X,sensitivity,MM,aniso)

Arguments

`copula`	String; the type of copula. It can be "Clayton" or "Gaussian"
`bivariate`	Logical; if `TRUE` then the data come froma a bivariate random field. Otherwise from a univariate random field.
`coordx`	A numeric (`d \times 2`)-matrix (where `d` is the number of points) assigning 2-dimensions of coordinates or a numeric vector assigning 1-dimension of coordinates.
`coordy`	A numeric vector assigning 1-dimension of coordinates; `coordy` is interpreted only if `coordx` is a numeric vector otherwise it will be ignored.
`coordt`	A numeric vector assigning 1-dimension of temporal coordinates. Optional argument, the default is `NULL` then a spatial random field is expected.
`coordx_dyn`	A list of `m` numeric (`d_t \times 2`)-matrices containing dynamical (in time) spatial coordinates. Optional argument, the default is `NULL`
`corrmodel`	Numeric; the id of the correlation model.
`data`	A numeric vector or a (`n \times d`)-matrix or (`d \times d \times n`)-matrix of observations.
`distance`	String; the name of the spatial distance. The default is `Eucl`, the euclidean distance. See the Section Details.
`flagcorr`	A numeric vector of binary values denoting which paramerters of the correlation function will be estimated.
`flagnuis`	A numeric vector of binary values denoting which nuisance paramerters will be estimated.
`fixed`	A numeric vector of parameters that will be considered as known values.
`GPU`	Numeric; if `NULL` (the default) no GPU computation is performed.
`grid`	Logical; if `FALSE` (the default) the data are interpreted as a vector or a (`n \times d`)-matrix, instead if `TRUE` then (`d \times d \times n`)-matrix is considered.
`likelihood`	String; the configuration of the compositelikelihood, see `GeoFit`.
`local`	Numeric; number of local work-items of the GPU
`lower`	An optional named list giving the values for the lower bound of the space parameter when the optimizer is `L-BFGS-B` or `nlminb` or `optimize`. The names of the list must be the same of the names in the `start` list.
`model`	Numeric; the id value of the density associated to the likelihood objects.
`n`	Numeric; number of trials in a binomial random fields.
`namescorr`	String; the names of the correlation parameters.
`namesnuis`	String; the names of the nuisance parameters.
`namesparam`	String; the names of the parameters to be maximised.
`numparam`	Numeric; the number of parameters to be maximised.
`numparamcorr`	Numeric; the number of correlation parameters.
`optimizer`	String; the optimization algorithm (see `optim` for details). `Nelder-Mead` is the default. Other possible choices are `nlm`, `BFGS` `L-BFGS-B` and `nlminb`. In these last two cases upper and lower bounds can be passed by the user. In the case of one-dimensional optimization, the function `optimize` is used.
`onlyvar`	Logical; if `TRUE` (and varest is `TRUE`) only the variance covariance matrix is computed without optimizing. `FALSE` is the default.
`parallel`	Logical; if `TRUE` optmization is performed using optimParallel using the maximum number of cores, when optimizer is `L-BFGS-B`.`FALSE` is the default.
`param`	A numeric vector of parameters' values.
`spacetime`	Logical; if `TRUE` the random field is spatial-temporal otherwise is a spatial field.
`type`	String; the type of the likelihood objects. If `Pairwise` (the default) then the marginal composite likelihood is formed by pairwise marginal likelihoods.
`upper`	An optional named list giving the values for the upper bound of the space parameter when the optimizer is or `L-BFGS-B` or `nlminb` or `optimize`. The names of the list must be the same of the names in the `start` list.
`varest`	Logical; if `TRUE` the estimate' variances and standard errors are returned. `FALSE` is the default.
`vartype`	String; the type of estimation method for computing the estimate variances, see `GeoFit`.
`weigthed`	Logical; if `TRUE` then decreasing weigths coming from a compactly supported correlation function with compact support `maxdist` (`maxtime`)are used.
`winconst`	Numeric; a positive value for computing the spatial sub-window in the sub-sampling procedure.
`winstp`	Numeric; a value in `(0,1]` for defining the the proportion of overlapping in the spatial sub-sampling procedure.
`winconst_t`	Numeric; a positive value for computing the temporal sub-window in the sub-sampling procedure.
`winstp_t`	Numeric; a value in `(0,1]` for defining the the proportion of overlapping in the temporal sub-sampling procedure.
`ns`	Numeric; Number of (dynamical) temporal instants.
`X`	Numeric; Matrix of space-time covariates in the linear mean specification.
`sensitivity`	Logical; if `TRUE` then the sensitivy matrix is computed
`MM`	Numeric;a non constant fixed mean
`aniso`	Logical; should anisotropy be considered?

Details

Subroutine called by GeoFit. The procedure estimates the model parameters by maximisation of the composite log-likelihood

Value

Return a list from an optim call.

Author(s)

Moreno Bevilacqua, moreno.bevilacqua89@gmail.com,https://sites.google.com/view/moreno-bevilacqua/home, Víctor Morales Oñate, victor.morales@uv.cl, https://sites.google.com/site/moralesonatevictor/, Christian", Caamaño-Carrillo, chcaaman@ubiobio.cl,https://www.researchgate.net/profile/Christian-Caamano