AICweights: Calculate Akaike information criterion model weights

In momentuHMM: Maximum Likelihood Analysis of Animal Movement Behavior Using Multivariate Hidden Markov Models

Calculate Akaike information criterion model weights

Description

Calculate Akaike information criterion model weights

Usage

AICweights(..., k = 2, n = NULL)

Arguments

...	momentuHMM, HMMfits, or miHMM objects, to compare AIC weights of the different models. The first object must be a momentuHMM, HMMfits, or miHMM object, but additional model objects can be passed as a list using the !!! operator (see rlang).
k	Penalty per parameter. Default: 2 ; for classical AIC.
n	Optional sample size. If specified, the small sample correction AIC is used (i.e., AICc = AIC + kp(p+1)/(n-p-1) where p is the number of parameters).

Details

• Model objects must all be either of class momentuHMM or multiple imputation model objects (of class HMMfits and/or miHMM).

• AIC is only valid for comparing models fitted to the same data. The data for each model fit must therefore be identical. For multiple imputation model objects, respective model fits must have identical data.

Value

The AIC weights of the models. If multiple imputation objects are provided, then the mean model weights (and standard deviations) are provided.

Examples

## Not run: 
# HMM specifications
nbStates <- 2
stepDist <- "gamma"
angleDist <- "vm"
mu0 <- c(20,70)
sigma0 <- c(10,30)
kappa0 <- c(1,1)
stepPar0 <- c(mu0,sigma0)
anglePar0 <- c(-pi/2,pi/2,kappa0)
formula <- ~cov1+cov2
                
# example$m is a momentuHMM object (as returned by fitHMM), automatically loaded with the package
mod1 <- fitHMM(example$m$data,nbStates=nbStates,dist=list(step=stepDist,angle=angleDist),
                Par0=list(step=stepPar0,angle=anglePar0),
                formula=~1,estAngleMean=list(angle=TRUE))

Par0 <- getPar0(mod1,formula=formula)                 
mod2 <- fitHMM(example$m$data,nbStates=nbStates,dist=list(step=stepDist,angle=angleDist),
                Par0=Par0$Par,beta0=Par0$beta,
                formula=formula,estAngleMean=list(angle=TRUE))
                
AICweights(mod1,mod2)

Par0nA <- getPar0(mod1,estAngleMean=list(angle=FALSE))                 
mod3 <- fitHMM(example$m$data,nbStates=nbStates,dist=list(step=stepDist,angle=angleDist),
                Par0=Par0nA$Par,beta0=Par0nA$beta,
                formula=~1)
 
AICweights(mod1,mod2,mod3)

# add'l models provided as a list using the !!! operator                               
AICweights(mod1, !!!list(mod2,mod3))

## End(Not run)

momentuHMM documentation built on Oct. 19, 2022, 1:07 a.m.