Package index • skmle

Package overview

skmle-package: skmle: Sieve Maximum Kernel-Weighted Log-Likelihood Estimation

Survival models with sparse longitudinal covariates

skmle(): Fit a Transformed Hazards Model by SMKLE
skmle_cv() print(<cv.skmle>): Select the Bandwidth by Cross-Validation
kee_cox(): Fit a Cox-Type KEE Model
kee_additive(): Fit an Additive Hazards KEE Model

Asynchronous longitudinal regression

Kernel-weighted estimating equations of Cao, Zeng and Fine (2015) for a response and a covariate observed on different time grids.

kee_async(): Asynchronous longitudinal regression with time-invariant coefficients
kee_async_cv() print(<cv.kee_async>): Choose the bandwidth for an asynchronous longitudinal fit
kee_async_td() print(<kee_td>): Asynchronous longitudinal regression with time-dependent coefficients

Bandwidth selection with bias correction

Fit an estimator over a grid of bandwidths, choose one by a plug-in mean squared error, extrapolate the coefficients to bandwidth zero, and report a standard error that accounts for the choice.

print(<bc>) summary(<bc>) print(<summary.bc>) skmle_bc(): Bias-Corrected SMKLE by Extrapolation in the Bandwidth
kee_cox_bc() kee_async_bc(): Bias-Corrected Kernel Estimating Equations by Extrapolation in the Bandwidth

Kernel weights

The weight a model gives a covariate observation at a given lag. Both model families take one through their weight argument, and any function of the standardised lag will do; the two Epanechnikov kernels here are the supplied starting points.

w_epan_half() w_epan_full(): Kernel weight functions

Simulation

sim_skmle_data(): Simulate Sparse Longitudinal Survival Data
sim_async_data(): Simulate asynchronous longitudinal data

Model summaries and plots

plot(<skmle>): Plot the estimated baseline function for skmle model
plot(<kee_td>): Plot estimated coefficient curves
vcov(<skmle>) vcov(<kee>) vcov(<kee_td>): Extract the covariance matrix of a fitted model
nobs(<skmle>) nobs(<kee>) nobs(<kee_td>): Number of subjects contributing to a fit
confint(<kee_td>): Wald confidence intervals for a coefficient curve
plot(<cv.kee_async>) plot(<cv.skmle>): Plot a cross-validation curve

Tidy output

broom generics returning tibbles, so fits compose with the rest of a tidy workflow.

tidy(<skmle>) tidy(<kee>) tidy(<kee_td>): Summarise a fit as a tibble
glance(<skmle>) glance(<kee>) glance(<kee_td>): One-row summary of a fit
augment(<kee_async>): Add fitted means to the covariate table
summary(<skmle>): Summary for skmle object
summary(<kee>): Summary for kee object
print(<skmle>): Print skmle object
print(<kee>): Print kee object
print(<summary.skmle>): Print summary of skmle object
print(<summary.kee>): Print summary of kee object