MOSS_hazard: onestep TMLE of treatment-rule specific survival curve

In wilsoncai1992/MOSS: One-Step TMLE for Survival Analysis

Description

updating the hazard using constrained step size update

Format

R6Class object.

Value

Object of R6Class with methods

Methods

iterate_onestep update the initial estimator

Public fields

A

vector of treatment

T_tilde

vector of last follow up time

Delta

vector of censoring indicator

density_failure

survival_curve object of predicted counterfactual survival curve

density_censor

survival_curve object of predicted counterfactual failure event survival curve

g1W

propensity score

A_intervene

the intervention of interest

k_grid

vector of interested time points

Methods

Public methods

• MOSS_hazard$new()

• MOSS_hazard$create_dNt()

• MOSS_hazard$construct_long_data()

• MOSS_hazard$fit_epsilon()

• MOSS_hazard$compute_mean_eic()

• MOSS_hazard$iterate_onestep()

• MOSS_hazard$clone()

---

Method new()

Usage

MOSS_hazard$new(
  A,
  T_tilde,
  Delta,
  density_failure,
  density_censor,
  g1W,
  A_intervene = NULL,
  k_grid = NULL
)

---

Method create_dNt()

Usage

MOSS_hazard$create_dNt()

---

Method construct_long_data()

Usage

MOSS_hazard$construct_long_data(
  A_intervene,
  density_failure,
  density_censor,
  which_A = "obs"
)

---

Method fit_epsilon()

Usage

MOSS_hazard$fit_epsilon(method = "l2", clipping = Inf)

---

Method compute_mean_eic()

Usage

MOSS_hazard$compute_mean_eic(psi_n, k_grid)

---

Method iterate_onestep()

Usage

MOSS_hazard$iterate_onestep(
  method = "l2",
  epsilon = 1,
  max_num_interation = 100,
  tmle_tolerance = NULL,
  verbose = FALSE
)

---

Method clone()

The objects of this class are cloneable with this method.

Usage

MOSS_hazard$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

  MOSS_hazard$new(
    A,
    T_tilde,
    Delta,
    density_failure,
    density_censor,
    g1W,
    A_intervene = 1,
    k_grid = 1:max(T_tilde)
  )

wilsoncai1992/MOSS documentation built on June 1, 2020, 2:26 p.m.