rm redundancy study_1.R

VK/simulation/study_1.R

@@ -199,204 +199,3 @@ save_results(simulation_results, filename)
 
 cat("Results saved to:", file.path(results_dir, filename), "\n")
 print(simulation_results$Summary)
-# study_1.R
-
-# Load necessary scripts and libraries
-source("gen_pop_data.R")
-source("calc_metrics.R")  # Ensure this file contains the updated metric calculation functions
-source("run_analysis.R")  # Source the helper functions
-library(furrr)
-library(parallel)
-library(dplyr)
-library(purrr)
-library(tidyr)
-library(progress)
-library(lavaan)
-
-# Generate seeds
-parallel_seeds <- function(n, seed = NULL) {
-  if (is.null(seed))
-    stop("seed must be provided.")
-  RNGkind("L'Ecuyer-CMRG")
-  set.seed(seed)
-  purrr::accumulate(seq_len(n - 1), function(s, x) parallel::nextRNGStream(s), 
-                    .init = .Random.seed)
-}
-
-# Generate parameters grid with seeds
-n_reps <- 10000
-params <- expand.grid(
-  seed = parallel_seeds(n_reps, seed = 42), # seed position in grid is essential or reorder results later - implicit repetition parameter
-  model_type = c("1.1", "1.2", "1.3", "1.4"),
-  N = c(100, 400, 6400),
-  reliability = c(0.3, 0.5, 0.7),
-  method = c("SEM", "gSAM", "lSAM_ML", "lSAM_ULS")
-)
-
-
-# Set population values
-B_true <- c(
-  'f3~f1' = 0.1, 'f3~f2' = 0.1, 'f3~f4' = 0.1,
-  'f4~f1' = 0.1, 'f4~f2' = 0.1,
-  'f5~f3' = 0.1, 'f5~f4' = 0.1
-)
-true_values <- list(
-  B = B_true
-)
-
-model_syntax_study1 <- "    
-    f1 =~ y1 + y2 + y3
-    f2 =~ y4 + y5 + y6
-    f3 =~ y7 + y8 + y9
-    f4 =~ y10 + y11 + y12
-    f5 =~ y13 + y14 + y15
-    
-    f3 ~ f1 + f2 + f4
-    f4 ~ f1 + f2
-    f5 ~ f3 + f4 + f1
-                        "
-# Study function
-simulate_inner <- function(model_type, N, reliability, method) {
-
-  safe_quiet_run_analysis <- safely(quietly(run_analysis))
-
-  data <- gen_pop_model_data(model_type, N, reliability)$data
-
-  fit_result <- safe_quiet_run_analysis(data, model_syntax_study1, method)
-
-  sanity_check_estimates <- run_sanity_check(model_type, model_syntax_study1)
-  sanity_check_results <- check_sanity(sanity_check_estimates, true_values)
-
-  warnings_detected <- fit_result$result$warnings
-  improper_solution <- any(grepl("some estimated ov variances are negative", warnings_detected))
-
-  if (!is.null(fit_result$result$result) && lavInspect(fit_result$result$result, "converged")) {
-    PT <- parTable(fit_result$result$result)
-    estimated_paths <- PT[PT$op == "~", "est"]
-    names(estimated_paths) <- paste0(PT[PT$op == "~", "lhs"], "~", PT[PT$op == "~", "rhs"])
-
-    # Calculate performance metrics
-    coverage <- calculate_coverage(fit_result$result$result, true_values)
-    relative_bias <- calculate_relative_bias(estimated_paths, true_values)
-    relative_rmse <- calculate_relative_rmse(estimated_paths, true_values)
-
-    list(
-      Converged = 1, NonConverged = 0,
-      EstimatedPaths = list(estimated_paths),
-      SanityCheck = list(sanity_check_estimates),
-      MaxSanityCheckDifference = sanity_check_results$MaxDifference,
-      SanityCheckAlarm = sanity_check_results$Alarm,
-      Coverage = coverage,
-      RelativeBias = relative_bias,
-      RelativeRMSE = relative_rmse,
-      RelativeBiasList = list(relative_bias),
-      RelativeRMSEList = list(relative_rmse),
-      ImproperSolution = improper_solution,
-      Warnings = toString(fit_result$result$warnings),
-      Messages = toString(fit_result$result$messages),
-      Errors = if (is.null(fit_result$error)) NA_character_ else toString(fit_result$error$message)
-    )
-  } else {
-    list(
-      Converged = 0, NonConverged = 1,
-      EstimatedPaths = list(setNames(rep(NA, length(true_values$B)), names(true_values$B))),
-      SanityCheck = list(sanity_check_estimates),
-      MaxSanityCheckDifference = sanity_check_results$MaxDifference,
-      SanityCheckAlarm = sanity_check_results$Alarm,
-      Coverage = NA,
-      RelativeBias = NA,
-      RelativeRMSE = NA,
-      RelativeBiasList = list(NA),
-      RelativeRMSEList = list(NA),
-      ImproperSolution = improper_solution,
-      Warnings = toString(fit_result$result$warnings),
-      Messages = toString(fit_result$result$messages),
-      Errors = if (is.null(fit_result$error)) NA_character_ else toString(fit_result$error$message)
-    )
-  }
-}
-
-simulate_mid <- function(seed, design) {
-  future_pmap(design, simulate_inner,
-              .options = furrr_options(seed = rep.int(list(seed), nrow(design))))
-}
-
-simulate_outer <- function(chunk, chunk_params) {
-  future_pmap(chunk_params, simulate_mid,
-              .options = furrr_options(seed = NULL, scheduling = 1))
-}
-
-run_study_1 <- function(params, true_values) {
-
-  # Run the simulations and analysis in parallel
-  results_df <- params %>%
-    mutate(chunk = row_number() %% 100) %>%
-    nest(.by = c(chunk, seed), .key = 'design') %>%
-    nest(.by = chunk, .key = 'chunk_params') %>%
-    mutate(results = future_map2(chunk, chunk_params, simulate_outer, .options = furrr_options(seed=NULL))) %>%
-    unnest(c(chunk_params, results)) %>%
-    unnest(c(design, results)) %>%
-    mutate(
-      Converged = map_dbl(results, ~ .x$Converged),
-      NonConverged = map_dbl(results, ~ .x$NonConverged),
-      Warnings = map_chr(results, ~ .x$Warnings),
-      Messages = map_chr(results, ~ .x$Messages),
-      Errors = map_chr(results, ~ .x$Errors),
-      EstimatedPaths = map(results, ~ .x$EstimatedPaths[[1]]),
-      SanityCheck = map(results, ~ .x$SanityCheck[[1]]),
-      MaxSanityCheckDifference = map_dbl(results, ~ .x$MaxSanityCheckDifference),
-      SanityCheckAlarm = map_lgl(results, ~ .x$SanityCheckAlarm),
-      Coverage = map_dbl(results, ~ .x$Coverage),
-      RelativeBias = map_dbl(results, ~ .x$RelativeBias),
-      RelativeRMSE = map_dbl(results, ~ .x$RelativeRMSE),
-      RelativeBiasList = map(results, ~ .x$RelativeBiasList[[1]]),
-      RelativeRMSEList = map(results, ~ .x$RelativeRMSEList[[1]]),
-      ImproperSolution = map_lgl(results, ~ .x$ImproperSolution)
-    )
-
-  print('Parallel computation done')
-  browser()
-
-
-  # Remove NAs from the lists before calculating MCSE
-  relative_bias_list <- unlist(results_df$RelativeBiasList)
-  relative_rmse_list <- unlist(results_df$RelativeRMSEList)
-
-  relative_bias_list <- na.omit(relative_bias_list)
-  relative_rmse_list <- na.omit(relative_rmse_list)
-
-  # Calculate summary statistics
-  summary_stats <- results_df %>%
-    group_by(model_type, N, reliability, method) %>%
-    summarise(
-      ConvergenceRate = mean(Converged),
-      NonConvergenceCount = sum(NonConverged),
-      n_converged = sum(Converged),
-      MeanMaxSanityCheckDifference = mean(MaxSanityCheckDifference, na.rm = TRUE),
-      SanityCheckAlarmCount = sum(SanityCheckAlarm, na.rm = TRUE),
-      MeanCoverage = mean(Coverage, na.rm = TRUE),
-      MeanRelativeBias = mean(RelativeBias, na.rm = TRUE),
-      MeanRelativeRMSE = mean(RelativeRMSE, na.rm = TRUE),
-      MCSE_RelativeBias = calculate_mcse_bias(relative_bias_list),
-      MCSE_RelativeRMSE = calculate_mcse_rmse(relative_rmse_list),
-      ImproperSolutionsCount = sum(ImproperSolution, na.rm = TRUE),  # Update to count
-      .groups = 'drop'
-    ) %>%
-    arrange(model_type, N, reliability, method)
-
-  # Return summary statistics and detailed results
-  list(Summary = summary_stats, DetailedResults = results_df)
-}
-
-# Run complete simulation study
-cat("Starting simulation study 1...\n")
-simulation_results <- run_study_1(params, true_values)
-cat("Simulation study completed. Saving results...\n")
-
-# Save with timestamp
-timestamp <- format(Sys.time(), "%Y%m%d%H%M%S")
-filename <- paste0("simulation_results_study1r", timestamp, ".rda")
-save_results(simulation_results, filename)
-
-cat("Results saved to:", file.path(results_dir, filename), "\n")
-print(simulation_results$Summary)