Implement a 'G-statistic' Likelihood Ratio Test, similar to the Chi-Squared Independence Test.

G-Statistic_Likelihood_Ratio_Test.R

g_statistic_lr_test <- function(x) {
  
  chisq_results <- chisq.test(x)
  
  expected_counts <- chisq_results[['expected']]
  observed_counts <- chisq_results[['observed']]
  
  G_Statistic <- 2 * sum(observed_counts * log(observed_counts/expected_counts),
                         na.rm = TRUE)
  
  degrees_of_freedom <- chisq_results[['parameter']]
  
  p_value <- pchisq(q = G_Statistic,
                    df = degrees_of_freedom,
                    ncp = 0,
                    lower.tail = TRUE)
  
  return(list('statistic' = G_Statistic,
              'degrees_of_freedom' = degrees_of_freedom,
              'p_value' = p_value))
}

Check and make sure this is using the appropriate tail for the kinds of tests this will be used for.