Simulate a longitudinal panel and return a tidy long-format data.frame

Usage

morie_simulate_longitudinal_panel(
  n_individuals = 50,
  n_timepoints = 20,
  p_variables = 3,
  cov_kernel = "ar1",
  cov_rho = 0.5,
  ar_lags = 1L,
  ar_spectral_radius = 0.8,
  ar_decay = 0.6,
  missing_fraction = 0,
  outlier_fraction = 0,
  outlier_scale = 5,
  seed = 42L
)

Arguments

n_individuals: Number of subjects.
n_timepoints: Number of time-points per subject.
p_variables: Number of variables.
cov_kernel: Innovation covariance kernel.
cov_rho: Correlation parameter.
ar_lags: VAR lag order.
ar_spectral_radius: Target spectral radius (per lag).
ar_decay: Geometric decay across lags.
missing_fraction: Probability of NA mask per entry.
outlier_fraction: Probability of outlier amplification.
outlier_scale: Multiplicative factor for outliers.
seed: Non-negative integer seed.

Value

A data.frame with columns subject_id, t, variable, value.

Examples

if (FALSE) {
  df <- morie_simulate_longitudinal_panel(
    n_individuals = 30, n_timepoints = 10, p_variables = 4
  )
  head(df)
}