OUModel-class: Class "OUModel"
In smd: Sparse Multivariate Diffusions

Description Objects from the Class Slots Extends Methods Author(s) See Also Examples

The class OUModel represents multivariate Ornstein-Uhlenbeck diffusion models.

Objects can be created by calls of the form new("OUModel", ...).

A p-dimensional Ornstein-Uhlenbeck diffusion model is determined by three parameters,

A:: A p-vector (the constant drift coefficient).
B:: A p x p matrix (the linear drift coefficient).
C:: A symmetric, positive semidefinite p x p matrix (the infinitesimal covariance matrix).

The diffusion is a solution to a SDE of the form dX_t = B(X_t-A) dt + D dW_t where W is a p-dimensional Brownian Motion and D satisfies D*t(D) = C.

The class extends MultDiffModel, a virtual class which represents multivariate diffusion models.

parameters:: A list which contains components A, B and C corresponding to the parameters of the model. A must be a numeric vector, B and C must be of class matrix.

Class MultDiffModel (directly).

gradient: signature(object = "OUModel", parameters = NULL, lossType = 1, useSufficientStat = TRUE, ...): Returns the gradient \nabla l(B,A) of the type 1 or 2 loss function l as a list with two elements A and B (the derivatives with respect to the parameters A and B, respectively). Uses data from object. If parameters are missing, the parameters of object are used. If parameters are not missing, they must be specified as a list of parameters which are compatible with object. lossType must be either 1 or 2.
validateParameters: signature(object = "OUModel", parameters): May be used to check that parameters, a set of parameters corresponding to an object of class OUModel, has the correct form.
parToList: signature(object = "OUModel", x): Converts the numeric vector x into a list of parameters A, B and C which are compatible with an object of class OUModel (by column).
parToNumeric: signature(object = "OUModel", parameters): Converts parameters, a list of parameters corresponding to an object of class OUModel, into a numeric vector. If parameters are not specified, the parameters of object are converted.