sem_psm: Fit Structural Equation Model of Pre-Spawning Mortality

In ebuhle/SEMPSM: Structural Equation Modeling for Pre-Spawning Mortality Using Stan

Description

Fit a hierarchical Bayesian structural equation model to data on pre-spawning mortality and landscape and climatic variables using Stan.

Usage

sem_psm(
  psm,
  X,
  normal_indx,
  gamma_indx,
  L = 1,
  fit = NULL,
  I0_Z = 1,
  I_su = 1,
  I_su_Z = 1,
  I_fa = 1,
  I_fa_Z = 1,
  I_fit = rep(1, nrow(psm)),
  I_lpd = rep(1, nrow(psm)),
  pars = c("a0", "A", "Z", "phi", "g_mu_X", "mu_b0", "b0_Z", "sigma_b0", "b0",
    "mu_b_su", "b_su_Z", "sigma_b_su", "b_su", "mu_b_fa", "b_fa_Z", "sigma_b_fa", "b_fa",
    "sigma_psm", "p_psm", "ll_psm"),
  chains = 3,
  iter = 12000,
  warmup = 2000,
  thin = 5,
  ...
)

Arguments

psm	A data frame with rows corresponding to observations, including columns named site A factor giving site names. ppt_su A numeric variable giving summer precipitation in mm. ppt_fa A numeric variable giving fall precipitation in mm. n An integer variable giving the number of carcasss sampled. n_psm An integer variable giving the number of pre-spawning mortalities in each sample.
X	A matrix with rows corresponding to sites and columns corresponding to landscape variables, to be modeled as either normal- or gamma-distributed.
normal_indx	An integer vector giving indices of columns of X to be modeled as normally distributed.
gamma_indx	An integer vector giving indices of columns of X to be modeled as gamma-distributed.
L	Integer giving the number of latent landscape factors to fit.
fit	Stanfit object as returned by a previous call to SEMPSM::sem_psm(). If not NULL, initial values will be drawn randomly from the posterior samples in fit using SEMPSM::stan_init_cv(); otherwise they will be generated using SEMPSM::stan_init().
I0_Z	Integer (0/1) indicating whether to include main effects of latent landscape factors on pre-spawning mortality risk.
I_su	Integer (0/1) indicating whether to include a main effect of summer precipitation on pre-spawning mortality risk.
I_su_Z	Integer (0/1) indicating whether to include interactions between summer precipitation and latent landscape factors on pre-spawning mortality risk.
I_fa	Integer (0/1) indicating whether to include a main effect of fall precipitation on pre-spawning mortality risk.
I_fa_Z	Integer (0/1) indicating whether to include interactions between fall precipitation and latent landscape factors on pre-spawning mortality risk.
I_fit	Integer vector (0/1) of length nrow(psm) indicating whether each pre-spawning mortality observation should be included in the likelihood. This may be useful for fitting some observations while simulating from the posterior predictive distribution for others.
I_lpd	Integer vector (0/1) of length nrow(psm) indicating whether to evaluate the log posterior predictive density for each observation.
pars	A character vector specifying parameters of interest to be saved.
chains	A positive integer specifying the number of Markov chains.
iter	A positive integer specifying the number of iterations for each chain (including warmup).
warmup	A positive integer specifying the number of warmup (aka burnin) iterations per chain.
thin	A positive integer specifying the period for saving samples.
...	Other arguments to be passed to rstan::sampling().

Value

An object of S4 class rstan::stanfit.

ebuhle/SEMPSM documentation built on Aug. 8, 2020, 4:05 a.m.