Skip to contents

Newton-Raphson algorithm is used to solve the estimating equation bar S_n (delta) = 0

Usage

twoarmglmcount.dr(
  y,
  x.cate,
  time,
  trt,
  ps,
  f1.predictor,
  f0.predictor,
  error.maxNR = 0.001,
  max.iterNR = 150,
  tune = c(0.5, 2)
)

Arguments

y

Observed outcome; vector of size n

x.cate

Matrix of p.cate baseline covariates; dimension n by p.cate

time

Log-transformed person-years of follow-up; vector of size n

trt

Treatment received; vector of size n units with treatment coded as 0/1

ps

Estimated propensity scores for all observations; vector of size n

f1.predictor

Initial predictions of the outcome (expected number of relapses for one unit of exposure time) conditioned on the covariates x for treatment group trt = 1; mu_1(x), step 1 in the two regression; vector of size n

f0.predictor

Initial predictions of the outcome (expected number of relapses for one unit of exposure time) conditioned on the covariates x for treatment group trt = 0; mu_0(x), step 1 in the two regression; vector of size n

error.maxNR

A numerical value > 0 indicating the minimum value of the mean absolute error in Newton Raphson algorithm. Used only if score.method = 'contrastReg'. Default is 0.001.

max.iterNR

A positive integer indicating the maximum number of iterations in the Newton Raphson algorithm. Used only if score.method = 'contrastReg'. Default is 150.

tune

A vector of 2 numerical values > 0 specifying tuning parameters for the Newton Raphson algorithm. tune[1] is the step size, tune[2] specifies a quantity to be added to diagonal of the slope matrix to prevent singularity. Used only if score.method = 'contrastReg'. Default is c(0.5, 2).

Value

coef: Doubly robust estimators of the regression coefficients delta_0; vector of size p + 1 (intercept included) vcov: Variance-covariance matrix of the estimated coefficient delta_0; matrix of size p + 1 by p + 1 converge: Indicator that the Newton Raphson algorithm converged for delta_0; boolean