Learn R Programming

HQM (version 1.0)

Superefficient Estimation of Future Conditional Hazards Based on Marker Information

Description

Provides a nonparametric smoothed kernel estimator for the future conditional hazard rate function when time-dependent covariates are present, a bandwidth selector for the estimator's implementation and pointwise and uniform confidence bands. Methods used in the package refer to Bagkavos, Isakson, Mammen, Nielsen and Proust-Lima (2025) .

Copy Link

Version

Install

install.packages('HQM')

Monthly Downloads

250

Version

1.0

License

GPL (>= 2)

Maintainer

Dimitrios Bagkavos

Last Published

February 1st, 2025

Functions in HQM (1.0)

h_xtll

Local linear future conditional hazard rate estimation at a single time point
dij

D matrix entries, used for the implementation of the local linear kernel
h_xt_vec

Hqm estimator on the marker grid
get_h_xll

Local linear future conditional hazard rate estimator
h_xt

Local constant future conditional hazard rate estimation at a single time point
get_alpha

Marker-only hazard rate
pbc2

Mayo Clinic Primary Biliary Cirrhosis Data
to_id

Event data frame
llweights

Local linear weight functions
llK_b

Local linear kernel
make_sf

Survival function from a hazard
prep_cv

Prepare for Cross validation bandwidth selection
prep_boot

Precomputation for wild bootstrap
Q1

Bandwidth selection score Q1
b_selection

Cross validation bandwidth selection
b_selection_prep_g

Preparations for bandwidth selection
Kernels

Classical (unmodified) kernel and related functionals
Conf_bands

Confidence bands
make_N, make_Ni, make_Y, make_Yi

Occurance and Exposure on grids
auc.hqm

AUC for the High Quality Marker estimator
R_K

Bandwidth selection score R
bs.hqm

Brier score for the High Quality Marker estimator
Epan

Epanechnikov kernel
dataset_split

Split dataset for K-fold cross validation
get_h_x

Local constant future conditional hazard rate estimator
g_xt

Computation of a key component for wild bootstrap
lin_interpolate

Linear interpolation