Nothing
#' Create a stationary Ornstein--Uhlenbeck model
#'
#' A Gaussian continuous-time model, simulated exactly at arbitrary
#' observation times without burn-in. With diagonal drift, component i has
#' autocorrelation `exp(-gap/tau[i])`.
#'
#' @param tau Finite positive numeric vector of persistence times. Supply
#' exactly one of `tau` and `drift`.
#' @param cov Stationary process covariance, not innovation covariance. Only
#' the lower triangle is used. Both `cov` and the driving covariance
#' `K %*% cov + cov %*% t(K)` must be positive semidefinite, where
#' `K = diag(1/tau)` for persistence-time initialization. Singular
#' covariances are allowed.
#' @param mu Optional constant mean vector, defaulting to zero.
#' @param drift Square matrix K in the drift term `-K %*% (X - mu)`.
#' Its eigenvalues must have strictly positive real parts.
#'
#' @details Over a gap of length d, the autoregressive coefficient matrix is
#' the matrix exponential \eqn{\exp(-Kd)}. With constant gaps, the sampled
#' process is a VAR(1); with unequal gaps, the matrix varies from step to step.
#'
#' @return A `VarmapackOUModel` object with fields `tau`, `drift`, `cov`, `mu`, `r`,
#' and method `sim(t, nrep = 1L, rng = NULL)`. The result has dimensions
#' `r` by `length(t)` by `nrep`, including for one-dimensional models.
#' The unused parameter field (`tau` or `drift`) is `NULL`.
#' Times must be finite and strictly increasing.
#' The first observation is drawn from `N(mu, cov)`. Supply a
#' `randompack::randompack_rng()` through `rng` to control the stream.
#' `acvf(lags)` returns Cov(X(t+h), X(t)) as an r by r by length(lags) array.
#' `specrad(gap)` returns the sampled VAR(1) spectral radius at a fixed gap.
#'
#' @examples
#' rng <- randompack::randompack_rng()
#' rng$seed(123)
#' t <- varmapack_gamma_times(100, shape = 16, rng = rng)
#' model <- varmapack_ou_model(c(1, 2), matrix(c(1, .5, .5, 1), 2))
#' X <- model$sim(t, nrep = 3, rng = rng)
#' @export
varmapack_ou_model <- function(tau = NULL, cov, mu = NULL, drift = NULL) {
VarmapackOUModel$new(tau, cov, mu, drift)
}
.vp_rng_pointer <- function(rng) {
if (is.null(rng)) return(NULL)
if (!inherits(rng, "RandompackRNG") || is.null(rng$ptr))
stop("rng must be a randompack_rng() object", call. = FALSE)
rng$ptr
}
.vp_lags <- function(lags) {
lags <- .vp_numeric_array(lags, "lags")
if (!is.null(dim(lags))) stop("lags must be a vector", call. = FALSE)
as.double(lags)
}
VarmapackOUModel <- R6::R6Class(
"VarmapackOUModel",
public = list(
tau = NULL, drift = NULL, cov = NULL, mu = NULL, r = NULL,
initialize = function(tau, cov, mu = NULL, drift = NULL) {
if (is.null(tau) == is.null(drift))
stop("specify exactly one of tau and drift", call. = FALSE)
if (is.null(drift)) {
tau <- .vp_numeric_array(tau, "tau")
if (!is.null(dim(tau)) || length(tau) < 1L || any(tau <= 0))
stop("tau must be a positive vector", call. = FALSE)
tau <- as.double(tau)
r <- length(tau)
} else {
drift <- .vp_numeric_array(drift, "drift")
if (!is.matrix(drift) || nrow(drift) < 1L || nrow(drift) != ncol(drift))
stop("drift must be a nonempty square matrix", call. = FALSE)
r <- nrow(drift)
drift <- matrix(as.double(drift), r, r)
}
if (!is.numeric(cov) || !is.matrix(cov) ||
nrow(cov) != r || ncol(cov) != r)
stop("cov must be an r by r matrix", call. = FALSE)
cov <- matrix(as.double(cov), r, r)
cov[upper.tri(cov)] <- t(cov)[upper.tri(cov)]
if (is.null(mu)) mu <- numeric(r)
mu <- .vp_numeric_array(mu, "mu")
if (!is.null(dim(mu)) || length(mu) != r)
stop("mu must be a vector of length r", call. = FALSE)
parameter <- if (is.null(drift)) tau else drift
.Call("varmapack_ou_validate_R", parameter, cov, as.double(mu),
PACKAGE = "varmapack")
self$tau <- tau
self$drift <- drift
self$cov <- cov
self$mu <- as.double(mu)
self$r <- r
},
specrad = function(gap) {
gap <- .vp_numeric_array(gap, "gap")
if (length(gap) != 1L || !is.null(dim(gap)))
stop("gap must be a real scalar", call. = FALSE)
parameter <- if (is.null(self$drift)) self$tau else self$drift
.Call("varmapack_ou_specrad_R", parameter, self$cov, self$mu, as.double(gap),
PACKAGE = "varmapack")
},
acvf = function(lags) {
parameter <- if (is.null(self$drift)) self$tau else self$drift
.Call("varmapack_ou_acvf_R", parameter, self$cov, self$mu, .vp_lags(lags),
PACKAGE = "varmapack")
},
sim = function(t, nrep = 1L, rng = NULL) {
t <- .vp_numeric_array(t, "t")
if (!is.null(dim(t)) || length(t) < 1L || any(diff(t) <= 0))
stop("t must be a nonempty strictly increasing vector", call. = FALSE)
nrep <- .vp_count(nrep, "nrep", 1L)
parameter <- if (is.null(self$drift)) self$tau else self$drift
.Call("varmapack_ou_sim_R", parameter, self$cov, self$mu, as.double(t),
nrep, .vp_rng_pointer(rng), PACKAGE = "varmapack")
}
)
)
#' Generate observation times with Gamma spacing
#'
#' @param n Number of times, at least one.
#' @param shape Positive Gamma shape. `Inf` gives constant spacing.
#' @param mean_gap Positive expected gap, in the same units as OU persistence.
#' @param rng Optional `randompack::randompack_rng()` object.
#'
#' @return A strictly increasing numeric vector starting at zero. The `n-1`
#' independent gaps have scale `mean_gap/shape` and coefficient of variation
#' `1/sqrt(shape)`. Their realized mean is not normalized. Unrepresentable
#' positive gaps or times produce an error.
#' @examples
#' t <- varmapack_gamma_times(100, shape = 4, mean_gap = 1)
#' @export
varmapack_gamma_times <- function(n, shape = 16, mean_gap = 1, rng = NULL) {
n <- .vp_count(n, "n", 1L)
if (!is.numeric(shape) || length(shape) != 1L || is.na(shape) || shape <= 0)
stop("shape must be positive or Inf", call. = FALSE)
if (!is.numeric(mean_gap) || length(mean_gap) != 1L ||
!is.finite(mean_gap) || mean_gap <= 0)
stop("mean_gap must be positive and finite", call. = FALSE)
.Call("varmapack_gamma_times_R", n, as.double(shape), as.double(mean_gap),
.vp_rng_pointer(rng), PACKAGE = "varmapack")
}
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.