coxph_mpl.object {survivalMPL}

MPL Proportional Hazards Regression Object

Description

This class of objects is returned by the coxph_mpl class of functions to represent a proportional hazards model fitted by maximum penalised likelihood. Objects of this class have methods for the functions print, summary, plot, residuals and predict.

All components described under Arguments must be included in a legitimate coxph_mpl object.

Arguments

coef

a list of length 2 containg the parameter estimates of each model part. The first list, named 'Beta', contains the vector of regression parameter estimates of length p. The second list, named 'Theta', contains the vector of the baseline hazard parameter estimates of length m.

se

a list of length 2 containg the parameter standard errors of each model part. The first list, named 'Beta', is a (p x 5) matrix indicating the standard errors of each regression parameter according to 5 inference methods. See Details in coxph_mpl. The second list, named 'Theta', is a (m x 5) matrix indicating the standard errors of each baseline hazard parameter according to 5 inference methods. See Details in coxph_mpl.

covar

a list of length 5 containg the (p+m x p+m) covariance matrix according to 5 inference methods. See Details in coxph_mpl.

ploglik

a vector of length 2. The first element is the penalised log-likelihood with the final values of the coefficients. (The second element is a correction factor for the baseline hazard parameters due to the use of a centered X matrix in the estimation process.)

iter

a vector of length 3 indicating the number of iterations used to estimate the smoothing parameter (first value, equal to 1 when the user specified a chosen value), the Beta and Theta parameters during the entire process (second value), and Beta and Theta parameters during the last smoothing parameter iteration (third value).

knots

list of length 3 to 4 containg parameters of the chosen basis: 'm', the number of used bases; 'Alpha', the knot sequence of length m+1 for the uniform basis, and of length m otherwise; 'Delta', the value of the integral of each base over the data support (which equals 1 when basis != 'uniform'); 'Sigma', only available for the Gaussian basis, corresponds to the standard deviation of each truncated Gaussian base.

control

Object of class coxph_mpl.control specifying the basis, smoothing parameter value and other options. See coxph_mpl.control.

dim

a list of length 5 with following elements: 'n', the sample size; 'n.events', the number of events; 'n.ties', the number of duplicated observations; 'p', the number of regression parameters; and 'm', the number of baseline hazard parameters.

call

the matched call.

data

a list of length 3 with following elements: 'time', the outcome vector with an added noise applied to duplicated observation if ties == "epsilon" in coxph_mpl.control; 'observed', a logical vector indicating if outcomes are fully observed or censored; 'X', the X matrix corresponding to the model formula indicated in coxph_mpl.

Author(s)

Dominique-Laurent Couturier, Maurizio Manuguerra

See Also

coxph_mpl, summary.coxph_mpl, coef.coxph_mpl, plot.coxph_mpl,residuals.coxph_mpl and predict.coxph_mpl.