Skip to contents

Log-likelihood for Gaussian AD models (with missing data support)

Source: R/loglik_gau.R
logL_gau.Rd

Computes the log-likelihood for Gaussian antedependence models of order 0, 1, or 2. Supports missing data under MAR assumption via na_action parameter.

Usage

logL_gau(
  y,
  order,
  mu = NULL,
  phi = NULL,
  sigma = NULL,
  blocks = NULL,
  tau = 0,
  na_action = c("fail", "complete", "marginalize")
)

Arguments

y

Numeric matrix with n_subjects rows and n_time columns. May contain NA.

order

Antedependence order, one of 0, 1, or 2.

mu

Mean vector (length n_time).

phi

Dependence coefficient(s). For order 1: vector of length n_time-1. For order 2: matrix with 2 columns or vector of length 2*(n_time-2).

sigma

Innovation standard deviations (length n_time).

blocks

Integer vector of block membership (length n_subjects), or NULL.

tau

Block effects, first element constrained to zero

na_action

How to handle missing values:

  • fail: Error if any NA is present (default)

  • complete: Use only complete cases

  • marginalize: Compute observed-data likelihood

Value

Scalar log-likelihood value.

Details

For complete data (no NA), all three na_action options give the same result.

For missing data:

  • marginalize: Uses MVN marginalization to compute P(Y_obs). This is the correct observed-data likelihood for MAR missing data.

  • complete: Removes subjects with any missing values. May lose information.

  • fail: Stops with error. Useful to ensure no missing data present.

Examples

set.seed(1)
y <- simulate_gau(n_subjects = 30, n_time = 5, order = 1, phi = 0.3)
fit <- fit_gau(y, order = 1)
logL_gau(y, order = 1, mu = fit$mu, phi = fit$phi, sigma = fit$sigma)
#> [1] -190.3954