dent_likelihood: Dents the likelihood surface This takes any values that are...

In OUwie: Analysis of Evolutionary Rates in an OU Framework

Dents the likelihood surface This takes any values that are better (lower) than the desired negative log likelihood and reflects them across the best_neglnL + delta line, "denting" the likelihood surface.

Description

Dents the likelihood surface This takes any values that are better (lower) than the desired negative log likelihood and reflects them across the best_neglnL + delta line, "denting" the likelihood surface.

Usage

dent_likelihood(neglnL, best_neglnL, delta = 2)

Arguments

neglnL	The original negative log likelihood
best_neglnL	The negative log likelihood at the optimum; other values will be greater than this.
delta	How far from the optimal negative log likelihood to focus samples

Value

The transformed negative log likelihood

Examples

sims <- stats::rnorm(100, mean=17)
possible_means <- seq(from=16, to=18, length.out=100)
results_normal <- rep(NA, length(possible_means))
results_dented <- results_normal
dnorm_to_run <- function(par, sims) {
  return(-sum(dnorm(x=sims, mean=par, log=TRUE)))
}
best_neglnL <- optimize(dnorm_to_run,interval=range(possible_means), 
  sims=sims, maximum=FALSE)$objective
for (i in seq_along(possible_means)) {
  results_normal[i] <- dnorm_to_run(possible_means[i], sims=sims)
  results_dented[i] <- dent_likelihood(results_normal[i], best_neglnL=best_neglnL, delta=2)
}

plot(possible_means, results_normal)
lines(possible_means, results_dented)

OUwie documentation built on Sept. 17, 2026, 1:09 a.m.