This function is used to statistical inference of model parameters. The methods including maximum likelihood estimator and Bayesian inference.
Usage
sta_infer(
method = "MLE",
process = "Wiener",
type = "classical",
data = NULL,
p = 0.975,
par = c(1, 1),
chains = 1,
warmup = 1000,
iter = 2000,
cores = 1,
s = NULL,
rel = NULL
)Arguments
- method
time.
- process
Wiener, Gamma or Inverse Gaussian process.
- type
classical in default.
- data
degradation data.
- p
in default `p=0.975`.
- par
initial parameter.
- chains
in default `chains = 1`
- warmup
in default `warmup = 1000`.
- iter
in default `iter= 2000`.
- cores
cores = 1,
- s
stress.
- rel
relationship.
Examples
dat <- sim_dat(
group = 5, t = 1:200, para = c(2, 3),
process = "Wiener", type = "classical",
s = NULL, rel = NULL
)
# MLE
mle_fit <- sta_infer(
method = "MLE", process = "Wiener",
type = "classical", data = dat
)
mle_fit
#> low mean up
#> [1,] 1.7401 1.9246 2.1092
#> [2,] 2.8396 2.9701 3.1006
# Bayes
# rstan_options(auto_write = TRUE)
# options(mc.cores = parallel::detectCores())
# bayes_fit = sta_infer(method = "Bayes", process = "Wiener",
# type = "classical", data = dat)
# bayes_fit$summary
# print(bayes_fit$summary, probs = c(0.025,0.5,0.975),pars = c("mu","w"))
# plot(bayes_fit$stan_re)
# traceplot(bayes_fit$stan_re,pars = c("mu","w"),
# inc_warmup = TRUE,nrow = 1) +
# theme(legend.position = "top")