AIC: Compute the Akaike Information Criterion (AIC) of a Gaussian mixture model or graphical model

Description

This function computes the Akaike Information Criterion (AIC) of a Gaussian mixture model or graphical model: $A I C = l o g L i k - n_{p a r}$ where $l o g L i k$ is the log-likelihood and $n_{p a r}$ the number of free parameters.

Usage

# S3 method for gmm
AIC(object, data, y = NULL, regul = 0.01, ...)
# S3 method for gmbn
AIC(object, data, col_seq = NULL, ...)
# S3 method for gmdbn
AIC(object, data, col_seq = NULL, ...)

Value

If object is a gmm object, a numeric value corresponding to the AIC.

If object is a gmbn or gmdbn object, a list with elements:

global: A numeric value corresponding to the global AIC.
local: For a gmbn object, a numeric vector containing the local conditional AICs. For a gmdbn object, a list of numeric vectors containing these values for each gmbn element.

Arguments

object: An object of class gmm, gmbn or gmdbn.
data: A data frame containing the data used to compute the AIC. Its columns must explicitly be named after the variables (or nodes) of object. If object is a gmm object, a numeric matrix can be passed.
y: A character vector containing the dependent variables if a conditional AIC is computed. If NULL (the default), the joint AIC is computed.
regul: A positive numeric value corresponding to the regularization constant if a penalty term is added for Bayesian regularization. If NULL, no penalty term is added. If a conditional AIC is computed, this argument is ignored.
...: Unused arguments from the generic function.
col_seq: A character vector containing the column names of data that describe the observation sequence. If NULL (the default), all the observations belong to a single sequence. If object is a temporal gmbn or gmdbn object, the observations of a same sequence must be ordered such that the $t$ th one is related to time slice $t$ (note that the sequences can have different lengths). If object is a non-temporal gmbn object, this argument is ignored.

Examples

Run this code

data(gmm_body, data_body)
aic_1 <- AIC(gmm_body, data_body)
aic_2 <- AIC(gmm_body, data_body, y = "WAIST")

data(gmbn_body, data_body)
aic_3 <- AIC(gmbn_body, data_body)

data(gmdbn_air, data_air)
aic_4 <- AIC(gmdbn_air, data_air, col_seq = "DATE")

Run the code above in your browser using DataLab