MQMF (version 0.1.1)

negNLL: negNLL -ve log-likelihood for normally distributed variables

Description

negNLL - Calculates the negative log-likelihood for normally distributed variables. It requires a function, 'funk' as an argument, that will calculate predicted values of a variable from a vector of input values. In the example below the predicted values are lengths-at-age and the input data are ages. Only the arguments used within negNLL are listed explicitly, which leaves the data required to drive the funk to generate the predicted values for comparison with the observed to be passed using the ..., so take care with spelling of the variable name required by whatever funk is being used.

Usage

negNLL(pars, funk, observed, ...)

Arguments

pars

a vector containing the parameters being used in funk, plus the sigma, which is the standard deviation of the normal random likelihoods in dnorm, an extra estimated parameter. Because negNLL refers explicitly to sigma, which must be the last parameter in the vector, it must be listed in the arguments.

funk

the function name that calculates the predicted values from the independent values passed using the ellipsis.

observed

the observed values of the variable that the model will predict to compare with each of the input observed values.

...

required to allow funk to access its other input data without having to explicitly define them in negNLL

Value

the sum of the negative log-likelihoods using a normal PDF

Examples

# NOT RUN {
 data(minnow)
 pars <- c(89, 0.01,-15,3.75)  # ssq    = 785.6183
 age <- minnow$week            # negNLL = 151.1713
 ssq(funk=vB,observed=minnow$length,p=pars,ages=age)
 negNLL(pars,funk=vB,observed=minnow$length,ages=age)
 oldpar <- par(no.readonly=TRUE)
 plot1(age,vB(pars,age),xlab="Age",ylab="Length",lwd=2)
 points(age,minnow$length,pch=16,cex=1.2)
 par(oldpar)
# }