stonegold546

library(lavaan)
library(rstan)
library(cmdstanr)

set_cmdstan_path("~/cmdstan/")

source("create_poly_stan_missing.R")

HS9 <- HolzingerSwineford1939[, paste0("x", 1:9)]
# Pearson correlations

stonegold546

library(rstan)

options(mc.cores = parallel::detectCores())
rstan_options(auto_write = TRUE)

# Simulate an example
set.seed(1234)
hist(x_true <- rt(2e2, 3) * 1.5)
# Create outcome
hist(y <- scale(rnorm(length(x_true), -.075 * x_true))[, 1])

stonegold546

functions {
  int sum0(vector x) {
    int len = num_elements(x);
    int res = 0;
    for (i in 1:len) if (x[i] == 0) res += 1;
    return res;
  }
}
data {
  int N;

stonegold546

library(rstan)
library(lavaan)

cmdstanr::set_cmdstan_path("~/cmdstan/")

# https://rfarouni.github.io/assets/projects/BayesianFactorAnalysis/BayesianFactorAnalysis.html
# L is N_items BY N_factor loading matrix, constrained lower-triangular
# Cov = LL' + diag(res_vars), factors constrained orthogonal

# Chose items 3, 6 and 8 as markers of factors 1, 2 and 3

stonegold546

library(data.table)
library(ggplot2)
library(scales)

set.seed(12345)
post_df <- data.table(
  est = rnorm(13 * 12 / 2, 0, .05),
  se = rlnorm(13 * 12 / 2, log(.09), log(.20 / .09) / qnorm(.9999))
)
post_df[, lo := qnorm(.025, est, se)]

stonegold546

### Block 02 -- Load packages

library(ordinalCont)  # Just to check frequentist
library(splines2)
library(brms)

### Block 03 -- Load helper functions

source("helper_file_ord_generic.R")