Title: Semi-Supervised Calibration of Risk with Noisy Event Times
Version: 0.1.1
Description: A consistent, semi-supervised, non-parametric survival curve estimator optimized for efficient use of Electronic Health Record (EHR) data with a limited number of current status labels. See van der Laan and Robins (1997) <doi:10.2307/2670119>.
License: GPL-3
Encoding: UTF-8
RoxygenNote: 7.1.1
Imports: Matrix, survival, pracma, foreach, doParallel, parallel, Rcpp
LinkingTo: Rcpp, RcppArmadillo
Suggests: knitr, rmarkdown
VignetteBuilder: knitr
LazyData: true
URL: https://github.com/celehs/SCORNET
BugReports: https://github.com/celehs/SCORNET/issues
NeedsCompilation: yes
Packaged: 2021-01-04 03:07:12 UTC; yuriahuja
Author: Yuri Ahuja [aut, cre]
Maintainer: Yuri Ahuja <Yuri_Ahuja@hms.harvard.edu>
Repository: CRAN
Date/Publication: 2021-01-04 20:40:03 UTC

SCORNET: A novel non-parametric survival curve estimator for the Electronic Health Record

Description

Semi-Supervised Calibration of Risk with Noisy Event Times (SCORNET) is a consistent, non-parametric survival curve estimator that boosts efficiency over existing non-parametric estimators by (1) utilizing unlabeled patients in a semi-supervised fashion, and (2) leveraging information-dense engineered EHR features to maximize unlabeled set imputation precision See Ahuja et al. (2020) BioArxiv for details

SCORNET Estimator

Description

SCORNET Estimator

Usage

scornet(
  Delta,
  C,
  t0.all,
  C.UL = NULL,
  filter = NULL,
  filter.UL = NULL,
  Z0 = NULL,
  Z0.UL = NULL,
  Zehr = NULL,
  Zehr.UL = NULL,
  K = Knorm,
  b = NULL,
  bexp = -1/4,
  fc = NULL,
  nCores = 1
)

Arguments

Delta

Labeled set current status labels (I(T<C))

C

Labeled set censoring times

t0.all

Times at which to estimate survival

C.UL

Unlabeled set censoring times

filter

Labeled set binary filter indicators

filter.UL

Unlabeled set filter indicators

Z0

Labeled set baseline feature matrix

Z0.UL

Unlabeled set baseline feature matrix

Zehr

Labeled set EHR-derived feature matrix

Zehr.UL

Unlabeled set EHR-derived feature matrix

K

Kernel function (defaults to standard normal)

b

bandwidth (optional)

bexp

bandwidth exponent (must be between -1/5 and -1/3, defaults to -1/4)

fc

N^1/4-consistent pdf estimator of C|Z0 (defaults to Kernel-Smoothed Cox/Breslow estimator)

nCores

Number of cores to use for parallelization (defaults to 1)

Value

S_hat: Survival function estimates at times t0.all; StdErrs: Asymptotically consistent standard error estimates corresponding to S_hat