| Dens1d.object {DALSM} | R Documentation |
Object created by Dens1d to prepare for density estimation from censored data using densityLPS
Description
An object returned by function Dens1d to prepare for density estimation with given mean and variance from censored data using function densityLPS.
Value
A Dens1d.object is a list with the following elements:
n: total sample size.y: a n-vector if no interval-censored data, a nx2 matrix otherwise (for interval-censored data: left and right limits of the interval ; for right-censored data: the right limit is set to Inf).event: a n-vector of observation indicators (0: right-censored ; 1: exactly observed or interval-censored).ymin: left limit of the variable support.ymax: right limit of the variable support.is.uncensored: boolean n-vector indicating if the corresponding <y> value is not right- or interval-censored.n.uncensored: total number of uncensored data.is.IC: boolean n-vector indicating if the correspondingyvalue is interval-censored.n.IC: total number of interval-censored data.is.RC: n-vector of booleans indicating if the correspondingyvalue is right-censored.n.RC: total number of right-censored data.ylow, yup: n-vector with the lower and upper limits of the interval data (when interval-censored). Whenyis exactly observed or right-censored, the two values are identical.ymid: n-vector containing the mean ofy.lowandy.up.K: number of B-splines in the basis used to model the log-hazard.knots: vector of knots for the B-spline basis.pen.order: penalty order.Pd: penalty matrix in the P-spline model.nbins: number of small bins used to partition the variable support.bins: small bins of equal width used to partition the variable support (cf. binning).ugrid: midpoints of the small bins.dbins: width of a small bin.Bbins: ((nbins +1) x K)-matrix containing the B-spline basis evaluated at the bin limits.C: (n x nbins) matrix of event or censoring indicatorsC_{ij}for unit 'i' and bin 'j'. For a unit with IC data, the bins with an non-empty intersection with the interval are indicated. When the unit is associated to a precise event time or censoring time in bin 'j', thenC_{ij}=1and 0 for other bins.Bgrid: (nbins x K)-matrix containing the B-spline basis evaluated at the bin midpoints.fgrid: nbins-vector with the estimated density values at the bin midpoints.rgrid: nbins-vector with the number of units 'still at risk' of an "event" in a given bin.BsB, ev: technical quantities used in the estimation procedure, see the code in Dens1d.R for more details.
Author(s)
Philippe Lambert p.lambert@uliege.be
References
Lambert, P. (2021). Fast Bayesian inference using Laplace approximations in nonparametric double additive location-scale models with right- and interval-censored data. Computational Statistics and Data Analysis, 161: 107250. <doi:10.1016/j.csda.2021.107250>
See Also
Dens1d, densityLPS, densLPS.object