TABLE OF CONTENTS

01structures/RCurve [ Definitions ]

NAME

    RCurve --- structure to store a curve in R

FUNCTION

This type of structure contains the information about a curve, either the hazard or frailty. This includes all parameters for spline and parametric components, weights, knot positions, etc, basis functions, tuning parameters, etc. It is the R analogue of the CCurve structure.

In R, this is implemented as a simple list. Most of the components are defined at initialization, others are added during the procedure.

The components are: 

        type                    "spline", "parametric" or "both"
        hasspline               boolean, whether a splien component is included
        haspar                  boolean, whether a parametric component is included
        name                    "hazard" or "frailty"
        spline.ord              order of the spline
        spline.adaptive         boolean, whether to use adaptive knot selection
        spline.knotspacing      "equal", "quantile", type of knot distribution 
        spline.nknots           number of knots
        spline.nknots.prior     prior on the number of knots
        spline.nknots.hyper     hyperparameters for the number of knots
        spline.ncandknots       number of candidate knots
        spline.candknots        candidate knots
        spline.maxoccknots      maximum number of occupied candidate knots
        spline.bdmconst         constant for birth-death-move step
        spline.knots            vector of knots
        spline.par              spline parameters
        spline.min              minimum value of a spline parameter
        spline.penalty          type of smoothness penalty to use ("none","2diff","2deriv")
        spline.penaltyfactor    scaling of the penalty
        spline.priorvar         prior variance for the spline penalty
        spline.hyper            hyperparameters for the prior variance
        spline.basis            B-spline basis
        spline.basisint         integrals of each B-spline basis function
        spline.basiscum         cumulative integrals of each basis function (hazard only)
        spline.basisexp         expectation of each spline basis function (frailty only)
        spline.candcov          covariance matrix used to generate candidate parameter sets
        spline.candcholcov      cholesky factorization of spline.candcov
        spline.fvar             frailty variance (frailty only)
        spline.fixedind         index of the spline component that remains fixed (frailty only)
        spline.tun              tuning parameter for the spline parameters
        spline.accept           acceptance rate of spline parameters
        param.dist              parametric distribution type
        param.par               parameters for the distribution
        param.priorvar          prior variance for the parametric parameters
        param.hyper             hyperparameters for the variance
        param.tun               tuning parameter
        param.accept            acceptance rate of parametric parameters
        weight                  weight of spline component
        weight.priorvar         variance of the weight
        weight.hyper            hyperparameters of the weight
        weight.tun              tuning parameter for the weight
        weight.accept           acceptance rate of the weight
        x                       observations (times for hazard, frailties for frailty curve)
        spline.y                spline component evaluated at x
        param.y                 parametric component evaluated at x
        y                       both components weighted
        spline.ycum             cumulative hazard evaluated at x using spline component
        param.ycum              cumulative hazard evaluated at y using parametric component
        ycum                    cumulative hazard