likelihood_fn: Likelihood function for the GrOU process with penalty.

In ntwk: Statistical Inference for Network Time Series

Description

Likelihood function for the GrOU process with penalty.

Usage

likelihood_fn(
  times,
  data,
  thresholds,
  lambda = NA,
  reg = "l1",
  div = 1e+05,
  gamma = NA,
  use_scaling = FALSE,
  mle = NA,
  log = TRUE
)

Arguments

times	Times at which data is given
data	Values to compute the MLE with.
thresholds	Jump threshold values.
lambda	Penalty parameter (defaults to NA with no penalty).
reg	Type of penalty (l1, l2 or adaptive).
div	Batch size/divisor to avoid large memory allocation.
gamma	Adaptive MLE scaling parameter.
use_scaling	Brownian motion covariance matrix scaling in the likelihood.
mle	MLE for adaptive regularisation.
log	Log-scale for the likelihood or not (defaults to FALSE).

Value

(Log)likelihood of the GrOU process with penalty.

Examples

n <- 1000
d <- 10
times <- seq(n)
delta_time <- 0.01
beta_value <- 0.499
noise <- matrix(rnorm(n * d, sd = sqrt(delta_time)), ncol = d)
data <- construct_path(
  diag(d),
  noise = noise, y_init = rep(0, d), delta_time = delta_time
)
thresholds <- rep(delta_time^beta_value, d)
loglik <- likelihood_fn(
  times = times, data = data,
  thresholds = thresholds,
  lambda = 1, div = 1e2
)
loglik(diag(d))

ntwk documentation built on Sept. 13, 2021, 9:07 a.m.