vgltsm.vgltsmff-class: Class of Vector Generalized Linear Time Series Models

In VGAMextra: Additions and Extensions of the 'VGAM' Package

Class of Vector Generalized Linear Time Series Models

Description

Time series family functions for the VGAMextra package

Objects from the Class

Objects can be created by calling new("vgltsmff"...)

slots

Implementation of vector generalized linear time series (TS) family functions (vgltsff) at VGAMextra is entirely based on the structure of family functions of the class vglmff-class.

Hence, refer to vglmff-class for a thourugh description of slots and features involved when objects of class "vgtsff" are being created.

Methods

Thus far, the following methods for objects of class "vgltsff-class" are implemented:

summary

Additional information to that displayed by the summary methods from VGAM. That is:

a) Standard errors based on the MLEs asymptotic distributions, and

b) Checks on stationarity and/or invertibility via the polynomial roots.

Currently, summary methods at VGAMextra have been implemented for:

signature(VGAMff = "ARff"):

For ARX–types family functions.

signature(VGAMff = "MAff"):

For MAX–types family functions.

signature(VGAMff = "ARMAff"):

For ARMAX–like family functions.

See summaryS4VGAMextra for further details.

Note

Programmers to write VGAM/VGLM time series family functions are also allowed to write methods functions either for specific purposes, or to extend those current methods to print some extra output required.

In such cases, notice that the class vgltsff-class is labeled by an object of class "character" (a character vector) specified at the slot @vfamily within the family function. This is, in fact, one of the required slots by the class vglmff-class.

Additionally, practitioners are encouraged to mantain all previous conventions for naming the arguments in Ts family functions as specified at vglmff-class, e.g., link is the argument for parameter link functions, etc.

Author(s)

V. Miranda and T.W. Yee.

VGAMextra documentation built on Nov. 2, 2023, 5:59 p.m.