Merge pull request #439 from spsanderson/development

Fixes #419

NAMESPACE

@@ -113,6 +113,7 @@ export(util_lognormal_param_estimate)
 export(util_lognormal_stats_tbl)
 export(util_negative_binomial_param_estimate)
 export(util_negative_binomial_stats_tbl)
+export(util_normal_aic)
 export(util_normal_param_estimate)
 export(util_normal_stats_tbl)
 export(util_pareto_param_estimate)

NEWS.md

@@ -9,7 +9,7 @@ None
 3. Fix #414 - Add function `util_chisquare_param_estimate()` to estimate the parameters of the chi-square distribution.
 4. Fix #417 - Add function `tidy_mcmc_sampling()` to sample from a distribution using MCMC.
 This outputs the function sampled data and a diagnostic plot.
-5. Fix #420 - Add functions `util_dist_aic()` functions to calculate the AIC for a distribution.
+5. Fix #421 - Add functions `util_dist_aic()` functions to calculate the AIC for a distribution.
 
 ## Minor Fixes and Improvements
 1. Fix #401 - Update `tidy_multi_single_dist()` to respect the `.return_tibble` parameter

R/utils-aic-normal.R

@@ -0,0 +1,58 @@
+#' Calculate Akaike Information Criterion (AIC) for Normal Distribution
+#'
+#' This function calculates the Akaike Information Criterion (AIC) for a normal distribution fitted to the provided data.
+#'
+#' @family Utility
+#' @author Steven P. Sanderson II, MPH
+#'
+#' @description
+#' This function estimates the parameters of a normal distribution from the provided data using maximum likelihood estimation,
+#' and then calculates the AIC value based on the fitted distribution.
+#'
+#' @param .x A numeric vector containing the data to be fitted to a normal distribution.
+#'
+#' @examples
+#' # Example 1: Calculate AIC for a sample dataset
+#' set.seed(123)
+#' data <- rnorm(30)
+#' util_normal_aic(data)
+#'
+#' @return
+#' The AIC value calculated based on the fitted normal distribution to the provided data.
+#'
+#' @name util_normal_aic
+#'
+#' @export
+#' @rdname util_normal_aic
+util_normal_aic <- function(.x) {
+  # Tidyeval
+  x <- as.numeric(.x)
+
+  # Get parameters
+  pe <- TidyDensity::util_normal_param_estimate(x)$parameter_tbl |> head(1)
+
+  # Negative log-likelihood function for normal distribution
+  neg_log_lik_norm <- function(par, data) {
+    mu <- par[1]
+    sigma <- par[2]
+    n <- length(data)
+    -sum(stats::dnorm(data, mean = mu, sd = sigma, log = TRUE))
+  }
+
+  # Fit normal distribution to population data (rnorm)
+  fit_norm <- optim(
+    c(pe$mu, pe$stan_dev),
+    neg_log_lik_norm,
+    data = x
+  )
+
+  # Extract log-likelihoods and number of parameters
+  logLik_norm <- -fit_norm$value
+  k_norm <- 2  # Number of parameters for normal distribution (mu and sigma)
+
+  # Calculate AIC
+  AIC_norm <- 2 * k_norm - 2 * logLik_norm
+
+  # Return
+  return(AIC_norm)
+}