minimizeme: The negative log likelihood for given parameters and data

In alko989/s6model: Data-poor stock assessment using a single-sepcies steady-state size-based (s6) model

The negative log likelihood for given parameters and data

Description

minimizeme is mainly used by an optimizer (e.g. optim) to estimate parameters.

Usage

minimizeme(theta, data, names, fixed.names = c(), fixed.vals = c(),
  isSurvey = FALSE)

minimizemeMultidata(theta, surdata, comdata, names, fixed.names = c(),
  fixed.vals = c())

Arguments

theta	Numeric vector of *transformed* parameter values
data	Numeric vector, or data.frame with columns named Weight and Freq. The observed values, fished individuals in grams
names	String vector. Contains the names of the parameter vector theta.
fixed.names	String vector. Names of constants.
fixed.vals	Numeric vector. Transformed values of constants.
isSurvey	Logical. If TRUE the observations are assumed to be from a survey.
surdata	Same as data. Survey data.
comdata	Same as data. Commercial data.

Value

Numeric scalar. The negative log likelihood for the given parameters and observations.

Note

If data is a list containing both sample and df, the data.frame df will be used.

Author(s)

alko

Examples

 
## Not run: 
## Simulate some data with default parameter values and fishing mortality Fm = 0.3
sim <- simulateData3(params = parameters("Fm", 0.3, transformed=FALSE))

## Plotting the negative log likelihood for different values of Fm
Fm <- seq(0.1, 1, 0.05) 
Fm.transformed <- log(Fm / 0.25)
nll <- sapply(Fm.transformed, function(x) minimizeme(theta = x, data = sim$sample, names = "Fm"))
plot(Fm, nll, type="l") 

## Using optimise to estimate one parameter
est <- optimise(f = minimizeme, interval=c(0.01, 1), data=sim$sample, names="Fm")
est.Fm <- exp(est$minimum) * 0.25
abline(v=est.Fm)
mtext(paste("Estimated Fm = ", round(est.Fm, 2)), at=est.Fm)

## End(Not run)

alko989/s6model documentation built on Nov. 2, 2023, 10:04 p.m.