Obtaining an optimized log likelihood value using 'fmincon'

According to the help file for 'aicbic' an optimized log likelihood can be obtained from 'fmincon', does this refer to the value of the objective function at the solution (i.e. fval)? If so, do I use fval or log(fval) as the input to 'aicbic'?

Thank you

Answer 1

[John D'Errico]

Just reading the help for this function, it does refer to the argument logL as being in a log form. I don't see where the confusion lies.

Description
example
aic = aicbic(logL,numParam) returns Akaike information criteria (AIC) corresponding to optimized loglikelihood function values (logL), as returned by estimate, and the model parameters, numParam.

Answer 2

[Star Strider]

This post fmincon and normal distribution log likelihood seems to bear directly on your Question.