lognorm_bstrp {marp}

A function to generate (double) bootstrap samples and fit Log-Normal renewal model

Description

A function to generate (double) bootstrap samples and fit Log-Normal renewal model

Usage

lognorm_bstrp(n, t, B, BB, par_hat, mu_hat, pr_hat, haz_hat, y)

Arguments

n

number of inter-event times

t

user-specified time intervals (used to compute hazard rate)

B

number of bootstrap samples

BB

number of double-bootstrap samples

par_hat

estimated parameters

mu_hat

estimated mean inter-event times

pr_hat

estimated time to event probability

haz_hat

estimated hazard rates

y

user-specified time point (used to compute time-to-event probability)

Value

returns list of estimates after fitting Log-Normal renewal model on (double) bootstrap samples

mu_star

Estimated mean from bootstrapped samples

pr_star

Estimated probability from bootstrapped samples

haz_star

Estimated hazard rates from bootstrapped samples

mu_var_hat

Variance of estimated mean

pr_var_hat

Variance of estimated probability

haz_var_hat

Variance of estimated hazard rates

mu_var_double

Variance of estimated mean of bootstrapped samples (via double-bootstrapping)

pr_var_double

Variance of estimated probability of bootstrapped samples (via double-bootstrapping)

haz_var_double

Variance of estimated hazard rates of bootstrapped samples (via double-bootstrapping)

mu_Tstar

Pivot quantity of the estimated mean

pr_Tstar

Pivot quantity of the estimated probability

haz_Tstar

Pivot quantity of the estimated hazard rates

Examples

# set some parameters
n <- 30 # sample size
t <- seq(100, 200, by = 10) # time intervals
B <- 100 # number of bootstraps
BB <- 100 # number of double-bootstraps
# m <- 10 # number of iterations for MLE optimization
par_hat <- c(
  3.41361e-03, 2.76268e+00, 2.60370e+00, 3.30802e+02, 5.48822e+00, 2.92945e+02, NA,
  9.43071e-03, 2.47598e+02, 1.80102e+00, 6.50845e-01, 7.18247e-01
)
mu_hat <- c(292.94512, 292.94513, 319.72017, 294.16945, 298.87286, 292.94512)
pr_hat <- c(0.60039, 0.42155, 0.53434, 0.30780, 0.56416, 0.61795)
haz_hat <-   matrix(c(
  -5.67999, -5.67999, -5.67999, -5.67999, -5.67999, -5.67999,
  -5.67999, -5.67999, -5.67999, -5.67999, -5.67999, -6.09420,
  -5.99679, -5.91174, -5.83682, -5.77031, -5.71085, -5.65738,
  -5.60904, -5.56512, -5.52504, -5.48833, -6.09902, -5.97017,
  -5.85769, -5.75939, -5.67350, -5.59856, -5.53336, -5.47683,
  -5.42805, -5.38621, -5.35060, -6.17146, -6.09512, -6.02542,
  -5.96131, -5.90194, -5.84668, -5.79498, -5.74642, -5.70064,
  -5.65733, -5.61624, -5.92355, -5.80239, -5.70475, -5.62524,
  -5.55994, -5.50595, -5.46106, -5.42359, -5.39222, -5.36591,
  -5.34383, -5.79111, -5.67660, -5.58924, -5.52166, -5.46879,
  -5.42707, -5.39394, -5.36751, -5.34637, -5.32946, -5.31596
),length(t),6)
y <- 304 # cut-off year for estimating probablity

# generate bootstrapped samples then fit renewal model
res <- marp::lognorm_bstrp(n, t, B, BB, par_hat, mu_hat, pr_hat, haz_hat, y)