spRR: Semiparametric targeted conditional relative risk/treatment...

In Larsvanderlaan/npOddsRatio: Semiparametric Targeted Maximum-Likelihood estimation of the conditional odds ratio in a partially-linear logistic-link regression model with post-treatment informative missingness.

Semiparametric targeted conditional relative risk/treatment effect using machine-learning RR(W) := E[Y|A=1/W] / E[Y|A=0,W]

Description

Semiparametric targeted conditional relative risk/treatment effect using machine-learning RR(W) := E[Y|A=1/W] / E[Y|A=0,W]

Usage

spRR(
  formula_logRR = ~1,
  W,
  A,
  Y,
  family_RR = gaussian(),
  sl3_Lrnr_A = NULL,
  sl3_Lrnr_Y = NULL,
  weights = NULL,
  smoothness_order = 1,
  max_degree = 2,
  num_knots = c(15, 5),
  fit_control = list()
)

Arguments

formula_logRR	R-formula object specifying model for log relative risk
W	A matrix of baseline covariates to condition on.
A	A binary treatment assignment vector
Y	A nonnegative outcome variable. Can be binary, a count, or a continuous nonnegative variable.
family_RR	A R-family object specifying the link function for the log relative risk (gaussian/identity implies formula_logRR is directly modelling the log RR)
sl3_Lrnr_A	An optional sl3-Learner object to estimate P(A=1|W)
sl3_Lrnr_Y	An optional sl3-Learner object to estimate nuisance conditional means E[Y|A=0,W] and E[Y|A=1,W]
weights	A vector of optional weights.
smoothness_order	Specification for default HAL learner (used if sl3 Learners not given). See spOR for use.
max_degree	Specification for default HAL learner (used if sl3 Learners not given). See spOR for use.
num_knots	Specification for default HAL learner (used if sl3 Learners not given). See spOR for use.
fit_control	Specification for default HAL learner (used if sl3 Learners not given). See spOR for use.

Larsvanderlaan/npOddsRatio documentation built on May 3, 2022, 12:05 p.m.