Nothing
R/bayesx.term.options.R
In R2BayesX: Estimate Structured Additive Regression Models with 'BayesX'
Defines functions
bxopts
Documented in
bxopts
bayesx.term.options <- bxopts <- function(bs = "ps", method = "MCMC")
{
bterms <- c("lasso", "ridge", "bl", "kr", "gk",
"gs", "mrf", "ps", "rw1", "rw2", "random", "te",
"season", "factor", "psplinerw1", "psplinerw2",
"geospline", "geokriging", "spatial", "kriging",
"baseline", "tensor")
if(!is.na(bs <- pmatch(bs, bterms))) {
bs <- bterms[bs]
cat("\nAvailable options for \'bs = \"", bs, "\"\': \n\n", sep = "")
if(bs %in% c("lasso", "ridge")) {
if(method == "MCMC") {
cat(" a, b: options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the shrinkage parameter lambda. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
}
if(method == "REML") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "bl" || bs == "baseline") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" gridchoice: How to choose grid points for numerical integration \n",
" in Cox and multistate models. May be either \"quantiles\", \n",
" \"equidistant\" or \"all\". \n",
" Default: \'gridchoice = \"quantiles\"\'. \n\n")
cat(" tgrid: number of equidistant time points to be used \n",
" for numerical integration in Cox and multi-state \n",
" models. Only meaningful if gridchoice = \"equidistant\". \n",
" Default: integer, \'tgrid = 100\'. \n\n")
cat("nrquantiles: number of quantiles that are used to define the \n",
" grid points for numerical integration in Cox and \n",
" multi-state models. First a grid of nrquantiles \n",
" quantiles is computed, then the grid for integration \n",
" is defined by nrbetween equidistant points between \n",
" each quantile. Only meaningful if \n",
" gridchoice = \"quantiles\". \n",
" Default: integer, \'nrquantiles = 50\'. \n\n")
cat(" nrbetween: number of points between quantiles that are used \n",
" to define the grid points for numerical integration \n",
" in Cox and multi-state models. First a grid of \n",
" nrquantiles quantiles is computed, then the grid for \n",
" integration is defined by nrbetween equidistant \n",
" points between each quantile. Only meaningful if \n",
" gridchoice = \"quantiles\". \n",
" Default: integer, \'nrbetween = 5\'. \n\n")
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "kr" || bs == "kriging") {
if(method == "REML") {
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" knotdata: a matrix with x and y coordinates for the knots. \n\n")
cat(" full: specifies that all distinct locations should be used as \n",
" knots in the kriging term. \n",
" Default: boolean, full = TRUE. \n\n")
cat(" nu: the smoothness parameter tau of the Matern correlation \n",
" function for kriging terms, admissible values are \n",
" 0.5, 1.5, 2.5, 3.5. \n",
" Default: \'nu = 1.5\'. \n\n")
cat(" maxdist: specifies the value c that is used to \n",
" determine the scale parameter p of the Matern correlation \n",
" function for kriging terms. \n",
" Default: realvalue, depends on nu, only positive. \n\n")
cat(" p, q: parameters of the coverage criterion for the space \n",
" filling algorithm that determines the knots of a kriging \n",
" term. \n",
" Default: realvalue, \'p = 20\', \'q = 20\'. \n\n")
cat(" maxsteps: maximum number of steps to be performed by the space \n",
" filling algorithm. \n",
" Default: integer, \'maxsteps = 300\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "gk" || bs == "geokriging") {
if(method == "REML") {
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" knotdata: a matrix with x and y coordinates for the knots. \n\n")
cat(" full: specifies that all distinct locations should be used as \n",
" knots in the kriging term. \n",
" Default: boolean, full = TRUE. \n\n")
cat(" nu: the smoothness parameter tau of the Matern correlation \n",
" function for kriging terms, admissible values are \n",
" 0.5, 1.5, 2.5, 3.5. \n",
" Default: \'nu = 1.5\'. \n\n")
cat(" maxdist: specifies the value c that is used to \n",
" determine the scale parameter p of the Matern correlation \n",
" function for kriging terms. \n",
" Default: realvalue, depends on nu, only positive. \n\n")
cat(" p, q: parameters of the coverage criterion for the space \n",
" filling algorithm that determines the knots of a kriging \n",
" term. \n",
" Default: realvalue, \'p = 20\', \'q = 20\'. \n\n")
cat(" maxsteps: maximum number of steps to be performed by the space \n",
" filling algorithm. \n",
" Default: integer, maxsteps = 300. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "te" || bs == "pspline2dimrw1" || bs == "tensor") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 8\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 8\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" gridsize: the option gridsize can be used to restrict the number of \n",
" points (on the x-axis) for which estimates are computed. By \n",
" default, estimates are computed at every distinct covariate \n",
" value in the data set (indicated by \'gridsize = -1\'). This \n",
" may be relatively time consuming in situations where the number \n",
" of distinct covariate values is large. If \'gridsize = 100\' \n",
" is specified, estimates are computed on an equidistant grid with \n",
" 100 knots. \n",
" Default: integer, \'gridsize = -1\'. \n\n")
}
if(method == "STEP") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" gridsize: may be used to restrict the number of points (on the x-axis) for \n",
" which estimates are computed. By default, estimates are computed at\n",
" every distinct covariate value in the data set - \'gridsize = -1\'.\n",
" This may be relatively time consuming in situations where the number\n",
" of distinct covariate values are large. If e.g. \'gridsize = 100\',\n",
" estimates are computed on an equidistant grid with 100 knots.\n",
" Default: integer, \'gridsize = -1\'.\n\n")
}
}
if(bs == "gs" || bs == "geospline") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\' \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
if(method == "STEP") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
}
if(bs == "random" || bs == "re" || bs == "ra") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\' \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" nofixed: the option nofixed suppresses the estimation of the main \n",
" effect for random slopes. \n",
" Default: boolean, \'nofixed = FALSE\'. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
}
}
if(bs == "mrf" || bs == "spatial") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
}
if(bs == "ps" || bs == "psplinerw1" || bs == "psplinerw2") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" gridsize: the option gridsize can be used to restrict the number of \n",
" points (on the x-axis) for which estimates are computed. By \n",
" default, estimates are computed at every distinct covariate \n",
" value in the data set (indicated by \'gridsize = -1\'). This \n",
" may be relatively time consuming in situations where the number \n",
" of distinct covariate values is large. If \'gridsize = 100\' \n",
" is specified, estimates are computed on an equidistant grid with \n",
" 100 knots. \n",
" Default: integer, \'gridsize = -1\'. \n\n")
cat(" derivative: if specified, first order derivatives of the function estimate \n",
" are computed. \n",
" Default: boolean, \'derivative = FALSE\'. \n\n")
cat(" monotone: defines monotonicity constraints for P-splines. Specifying \n",
" \'monotone = \"increasing\"\' yields increasing nonlinear \n",
" functions and \'monotone = \"decreasing\" yields decreasing \n",
" functions. \n",
" Default: \'monotone = \"unrestricted\"\'. \n\n")
cat("contourprob: forces the computation of contour probabilities for P-splines. \n",
" For instance, \'contourprob = 4\' specifies that contour \n",
" probabilities for difference orders zero to four are computed. \n",
" Note that the global simple option \'approx\' may additionally be \n",
" specified. In this case the computation of contour probabilities \n",
" is based on stochastic approximations for quantiles. \n",
" Default: \'contourprob = 4\'. \n\n")
}
if(method == "STEP") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
cat(" monotone: \'monotone = \"unrestricted\"\' - no constraint on the spline function,\n",
" \'monotone = \"increasing\"\' - monotonically increasing function,\n",
" \'monotone = \"decrasing\"\' - monotonically decreasing function,\n",
" \'monotone = \"convex\"\' - convex function, i.e. positive second\n",
" derivative,\n",
" \'monotone = \"concave\"\' - concave function\n",
" Default: \'monotone = \"unrestricted\"\'.\n\n")
cat(" gridsize: may be used to restrict the number of points (on the x-axis) for \n",
" which estimates are computed. By default, estimates are computed at\n",
" every distinct covariate value in the data set - \'gridsize = -1\'.\n",
" This may be relatively time consuming in situations where the number\n",
" of distinct covariate values are large. If e.g. \'gridsize = 100\',\n",
" estimates are computed on an equidistant grid with 100 knots.\n",
" Default: integer, \'gridsize = -1\'.\n\n")
}
}
if(bs == "season") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" period: period of a seasonal effect. The default \'period = 12\' \n",
" corresponds to monthly data. \n",
" Default: integer, \'period = 12\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" period: period of a seasonal effect. The default \'period = 12\' \n",
" corresponds to monthly data. \n",
" Default: integer, \'period = 12\'. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" period: period for the seasonal effect.\n",
" Default: numeric, \'period = 12\'.\n\n")
}
}
if(bs == "rw1" || bs == "rw2") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
}
}
if(bs == "factor") {
if(method == "REML") {
return(NULL)
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" coding: \'center = \"dummy\"\' - dummy coding of categorial variables,\n",
" \'center = effect\' - effect coding of categorial variables.\n",
" Default: character, \'coding = \"dummy\"\'.\n\n")
cat(" reference: specifies the reference category for categorial covariates.\n",
" Default: \'reference = 1\'.\n\n")
}
}
}
}
Try the R2BayesX package in your browser
Any scripts or data that you put into this service are public.
R2BayesX documentation built on Oct. 20, 2023, 3:01 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions bxopts
Documented in bxopts
bayesx.term.options <- bxopts <- function(bs = "ps", method = "MCMC")
{
bterms <- c("lasso", "ridge", "bl", "kr", "gk",
"gs", "mrf", "ps", "rw1", "rw2", "random", "te",
"season", "factor", "psplinerw1", "psplinerw2",
"geospline", "geokriging", "spatial", "kriging",
"baseline", "tensor")
if(!is.na(bs <- pmatch(bs, bterms))) {
bs <- bterms[bs]
cat("\nAvailable options for \'bs = \"", bs, "\"\': \n\n", sep = "")
if(bs %in% c("lasso", "ridge")) {
if(method == "MCMC") {
cat(" a, b: options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the shrinkage parameter lambda. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
}
if(method == "REML") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "bl" || bs == "baseline") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" gridchoice: How to choose grid points for numerical integration \n",
" in Cox and multistate models. May be either \"quantiles\", \n",
" \"equidistant\" or \"all\". \n",
" Default: \'gridchoice = \"quantiles\"\'. \n\n")
cat(" tgrid: number of equidistant time points to be used \n",
" for numerical integration in Cox and multi-state \n",
" models. Only meaningful if gridchoice = \"equidistant\". \n",
" Default: integer, \'tgrid = 100\'. \n\n")
cat("nrquantiles: number of quantiles that are used to define the \n",
" grid points for numerical integration in Cox and \n",
" multi-state models. First a grid of nrquantiles \n",
" quantiles is computed, then the grid for integration \n",
" is defined by nrbetween equidistant points between \n",
" each quantile. Only meaningful if \n",
" gridchoice = \"quantiles\". \n",
" Default: integer, \'nrquantiles = 50\'. \n\n")
cat(" nrbetween: number of points between quantiles that are used \n",
" to define the grid points for numerical integration \n",
" in Cox and multi-state models. First a grid of \n",
" nrquantiles quantiles is computed, then the grid for \n",
" integration is defined by nrbetween equidistant \n",
" points between each quantile. Only meaningful if \n",
" gridchoice = \"quantiles\". \n",
" Default: integer, \'nrbetween = 5\'. \n\n")
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "kr" || bs == "kriging") {
if(method == "REML") {
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" knotdata: a matrix with x and y coordinates for the knots. \n\n")
cat(" full: specifies that all distinct locations should be used as \n",
" knots in the kriging term. \n",
" Default: boolean, full = TRUE. \n\n")
cat(" nu: the smoothness parameter tau of the Matern correlation \n",
" function for kriging terms, admissible values are \n",
" 0.5, 1.5, 2.5, 3.5. \n",
" Default: \'nu = 1.5\'. \n\n")
cat(" maxdist: specifies the value c that is used to \n",
" determine the scale parameter p of the Matern correlation \n",
" function for kriging terms. \n",
" Default: realvalue, depends on nu, only positive. \n\n")
cat(" p, q: parameters of the coverage criterion for the space \n",
" filling algorithm that determines the knots of a kriging \n",
" term. \n",
" Default: realvalue, \'p = 20\', \'q = 20\'. \n\n")
cat(" maxsteps: maximum number of steps to be performed by the space \n",
" filling algorithm. \n",
" Default: integer, \'maxsteps = 300\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "gk" || bs == "geokriging") {
if(method == "REML") {
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" knotdata: a matrix with x and y coordinates for the knots. \n\n")
cat(" full: specifies that all distinct locations should be used as \n",
" knots in the kriging term. \n",
" Default: boolean, full = TRUE. \n\n")
cat(" nu: the smoothness parameter tau of the Matern correlation \n",
" function for kriging terms, admissible values are \n",
" 0.5, 1.5, 2.5, 3.5. \n",
" Default: \'nu = 1.5\'. \n\n")
cat(" maxdist: specifies the value c that is used to \n",
" determine the scale parameter p of the Matern correlation \n",
" function for kriging terms. \n",
" Default: realvalue, depends on nu, only positive. \n\n")
cat(" p, q: parameters of the coverage criterion for the space \n",
" filling algorithm that determines the knots of a kriging \n",
" term. \n",
" Default: realvalue, \'p = 20\', \'q = 20\'. \n\n")
cat(" maxsteps: maximum number of steps to be performed by the space \n",
" filling algorithm. \n",
" Default: integer, maxsteps = 300. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
return(NULL)
}
}
if(bs == "te" || bs == "pspline2dimrw1" || bs == "tensor") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 8\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 8\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" gridsize: the option gridsize can be used to restrict the number of \n",
" points (on the x-axis) for which estimates are computed. By \n",
" default, estimates are computed at every distinct covariate \n",
" value in the data set (indicated by \'gridsize = -1\'). This \n",
" may be relatively time consuming in situations where the number \n",
" of distinct covariate values is large. If \'gridsize = 100\' \n",
" is specified, estimates are computed on an equidistant grid with \n",
" 100 knots. \n",
" Default: integer, \'gridsize = -1\'. \n\n")
}
if(method == "STEP") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" gridsize: may be used to restrict the number of points (on the x-axis) for \n",
" which estimates are computed. By default, estimates are computed at\n",
" every distinct covariate value in the data set - \'gridsize = -1\'.\n",
" This may be relatively time consuming in situations where the number\n",
" of distinct covariate values are large. If e.g. \'gridsize = 100\',\n",
" estimates are computed on an equidistant grid with 100 knots.\n",
" Default: integer, \'gridsize = -1\'.\n\n")
}
}
if(bs == "gs" || bs == "geospline") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\' \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
if(method == "STEP") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
}
if(bs == "random" || bs == "re" || bs == "ra") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\' \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" nofixed: the option nofixed suppresses the estimation of the main \n",
" effect for random slopes. \n",
" Default: boolean, \'nofixed = FALSE\'. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
}
}
if(bs == "mrf" || bs == "spatial") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
cat(" map: a map provided as a list of matrix polygons. \n\n")
}
}
if(bs == "ps" || bs == "psplinerw1" || bs == "psplinerw2") {
if(method == "REML") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\'. \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" gridsize: the option gridsize can be used to restrict the number of \n",
" points (on the x-axis) for which estimates are computed. By \n",
" default, estimates are computed at every distinct covariate \n",
" value in the data set (indicated by \'gridsize = -1\'). This \n",
" may be relatively time consuming in situations where the number \n",
" of distinct covariate values is large. If \'gridsize = 100\' \n",
" is specified, estimates are computed on an equidistant grid with \n",
" 100 knots. \n",
" Default: integer, \'gridsize = -1\'. \n\n")
cat(" derivative: if specified, first order derivatives of the function estimate \n",
" are computed. \n",
" Default: boolean, \'derivative = FALSE\'. \n\n")
cat(" monotone: defines monotonicity constraints for P-splines. Specifying \n",
" \'monotone = \"increasing\"\' yields increasing nonlinear \n",
" functions and \'monotone = \"decreasing\" yields decreasing \n",
" functions. \n",
" Default: \'monotone = \"unrestricted\"\'. \n\n")
cat("contourprob: forces the computation of contour probabilities for P-splines. \n",
" For instance, \'contourprob = 4\' specifies that contour \n",
" probabilities for difference orders zero to four are computed. \n",
" Note that the global simple option \'approx\' may additionally be \n",
" specified. In this case the computation of contour probabilities \n",
" is based on stochastic approximations for quantiles. \n",
" Default: \'contourprob = 4\'. \n\n")
}
if(method == "STEP") {
cat(" degree: the degree of the B-spline basis functions.\n",
" Default: integer, \'degree = 3\'.\n\n")
cat(" knots: number of inner knots.\n",
" Default: integer, \'knots = 20\'.\n\n")
cat(" order: only if \'bs = \"ps\"\', the order of the difference penalty.\n",
" Default: integer, \'order = 2\'.\n\n")
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
cat(" monotone: \'monotone = \"unrestricted\"\' - no constraint on the spline function,\n",
" \'monotone = \"increasing\"\' - monotonically increasing function,\n",
" \'monotone = \"decrasing\"\' - monotonically decreasing function,\n",
" \'monotone = \"convex\"\' - convex function, i.e. positive second\n",
" derivative,\n",
" \'monotone = \"concave\"\' - concave function\n",
" Default: \'monotone = \"unrestricted\"\'.\n\n")
cat(" gridsize: may be used to restrict the number of points (on the x-axis) for \n",
" which estimates are computed. By default, estimates are computed at\n",
" every distinct covariate value in the data set - \'gridsize = -1\'.\n",
" This may be relatively time consuming in situations where the number\n",
" of distinct covariate values are large. If e.g. \'gridsize = 100\',\n",
" estimates are computed on an equidistant grid with 100 knots.\n",
" Default: integer, \'gridsize = -1\'.\n\n")
}
}
if(bs == "season") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat(" period: period of a seasonal effect. The default \'period = 12\' \n",
" corresponds to monthly data. \n",
" Default: integer, \'period = 12\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
cat(" period: period of a seasonal effect. The default \'period = 12\' \n",
" corresponds to monthly data. \n",
" Default: integer, \'period = 12\'. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" period: period for the seasonal effect.\n",
" Default: numeric, \'period = 12\'.\n\n")
}
}
if(bs == "rw1" || bs == "rw2") {
if(method == "REML") {
cat("lambdastart: starting value for the smoothing parameter lambda. \n",
" Default: realvalue, positive, \'lambdastart = 10\'. \n\n")
cat("catspecific: requests that the corresponding effect should be modelled \n",
" categoryspecific. Can only be used in cumulative and \n",
" sequential models for categorical responses, i.e. with \n",
" response families \"cumlogit\", \"cumprobit\", \"seqlogit\" \n",
" and \"seqprobit\". \n",
" Default: boolean, \'catspecific = FALSE\' \n\n")
cat(" center: for varying coefficient terms this option requests that \n",
" the effect should be centered to avoid identifiability \n",
" problems. \n",
" Default: boolean, \'center = FALSE\'. \n\n")
}
if(method == "MCMC") {
cat(" a, b: the options a and b specify the hyperparameters of the \n",
" inverse Gamma prior for the variance tau2. \n",
" Default: realvalue, \'a = 0.001\', \'b = 0.001\'. \n\n")
cat(" min, max: the options min and max define the minimum and maximum \n",
" block sizes of block move updates. In every iteration, \n",
" BayesX randomly chooses the block size within this range. \n",
" If omitted, the minimum and maximum block sizes are \n",
" automatically determined during the burnin period such that \n",
" the average acceptance rate is between 30 and 70 percent. \n",
" The specification of minimum and maximum block sizes is only \n",
" meaningful in combination with conditional prior proposals \n",
" and therefore has no effect for Gaussian responses, \n",
" categorical probit models, or if \'proposal = \"iwls\"\' or \n",
" \'proposal = \"iwlsmode\"\' is specified. \n",
" Default: automatic determination. \n\n")
cat(" lambda: provides a starting value for the smoothing parameter \n",
" lambda. \n",
" Default: realvalue, \'lambda = 0.1\'. \n\n")
cat(" proposal: specifies the type of proposal. \'proposal = \"cp\"\' means \n",
" conditional prior proposal, \'proposal = \"iwls\"\' stands for \n",
" iteratively weighted least squares (IWLS) proposal and \n",
" \'proposal = \"iwlsmode\"\' indicates IWLS based on posterior \n",
" mode estimation. \n",
" Default: \'proposal = \"iwls\"\'. \n\n")
cat(" updateW: the option updateW may be used to specify how often the IWLS \n",
" weight matrix should be updated. \'updateW = 0\' means never, \n",
" \'updateW = 1\' means in every iteration (which is the default), \n",
" \'updateW = 2\' means in every second iteration and so on. \n",
" Default: integer, \'updateW = 1\'. \n\n")
}
if(method == "STEP") {
cat(" dfmin: minimum degree of freedom, numeric.\n\n")
cat(" dfmax: maximum degree of freedom, numeric.\n\n")
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" logscale: \'logscale = FALSE\' - equidistant degrees of freedom,\n",
" \'logscale = TRUE\' - smoothing parameters on a logarithmic scale.\n",
" Default: \'logscale = FALSE\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spmin: minimum smoothing parameter.\n",
" Default: numeric, \'spmin = 1e-04\'.\n\n")
cat(" spmax: maximum smoothing parameter.\n",
" Default: numeric, \'spmax = 1e+04\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat(" number: number of different smoothing parameters.\n",
" Default: integer, \'number = 0\'.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" nofixed: \'nofixed = FALSE\' - linear fit is allowed,\n",
" \'nofixed = TRUE\' - linear fit is not allowed.\n",
" Default: boolean, \'nofixed = FALSE\'.\n\n")
cat(" center: \'center = FALSE\' - varying coefficient term is not centered,\n",
" \'center = TRUE\' - varying coefficient term is centered.\n",
" Default: boolean, \'center = FALSE\'.\n\n")
}
}
if(bs == "factor") {
if(method == "REML") {
return(NULL)
}
if(method == "MCMC") {
return(NULL)
}
if(method == "STEP") {
cat(" dfstart: degree of freedom used in the start model.\n",
" Default: numeric, \'dfstart = 1\'.\n\n")
cat(" sp: \'sp = \"automatic\"\' - list of smoothing parameters are \n",
" automatically specified. \'sp = \"df\"\' - smoothing parameters \n",
" are directly specified by the user in terms of degrees of freedom,\n",
" use options dfmin, dfmax, number, dfstart, logscale for specification.\n",
" \'sp = \"direct\"\' - smoothing parameters are directly by the user,\n",
" use options spmin, spmax, number, spstart, logscale for specification.\n",
" Default: character,\'sp = \"automatic\"\'.\n\n")
cat(" spstart: smoothing parameter for the start model, \'spstart = 0 excludes\n",
" the term from the start model, \'spstart = -1\' includes a linear\n",
" effect.\n\n")
cat("forced_into: \'forced_into = FALSE\' - term may be excluded from the model,\n",
" \'forced_into = TRUE\' - term may not be excluded from the model.\n",
" Default: boolean, \'forced_into = FALSE\'.\n\n")
cat(" coding: \'center = \"dummy\"\' - dummy coding of categorial variables,\n",
" \'center = effect\' - effect coding of categorial variables.\n",
" Default: character, \'coding = \"dummy\"\'.\n\n")
cat(" reference: specifies the reference category for categorial covariates.\n",
" Default: \'reference = 1\'.\n\n")
}
}
}
}
Try the R2BayesX package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.