grad_likelihood_fn: Gradient of the GrOU likelihood function with penalty.

In ntwk: Statistical Inference for Network Time Series

Description

Gradient of the GrOU likelihood function with penalty.

Usage

grad_likelihood_fn(
  times,
  data,
  thresholds,
  div = 1e+05,
  use_scaling = FALSE,
  log = TRUE
)

Arguments

times	Times at which data is given
data	Values to compute the MLE with.
thresholds	Jump threshold values.
div	Batch size/divisor to avoid large memory allocation.
use_scaling	Brownian motion covariance matrix scaling in the likelihood.
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)
grad_loglik <- grad_likelihood_fn(
  times = times, data = data, thresholds = thresholds, div = 1e2
)
grad_loglik(diag(d))

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