00Intro.dse1 {dse1} | R Documentation |
This is a library of functions for time series modeling. The library works with the S (Splus) and R languages. A "Brief User's Guide" is available at www.bank-banque-canada.ca/pgilbert This help and other information is also available at that web location.
The library implements an object oriented approach to time series modeling. This means that different model and data representations can be implemented with fairly simple extensions to the library.
The library includes multi-variate state space and ARMA (including VAR) models. The library also implements Troll models as an example of another class of model. (Troll models are run by completely separate software, Troll from Intex Inc, as if they were an integral part of the library. Models and data are passed back and forth to Troll and the results can be analyzed with already existing functions in the library. Note: the Troll interface is broken in the current version.)
The library includes methods for simulating, estimating, and converting among different model representations. These form the basic part of the library. Methods for studying estimation techniques and for examining the forecasting properties of models are also documented in the User's Guide.
There are also functions for forecasting and for evaluating the performance of forecasting models as well as functions for evaluating model estimation techniques.
DSE requires libraries tframe and setRNG. The first provides a kernel of functions for programming time series methods relatively independently of the representation of time. The second provides some utilities for setting and resetting the random number generator and normal transformations, and for generating the same random numbers in S and R differences. For programmers, these libraries may of interest by themselves. Tframe is intended to make it easier to write code which can use any new/better time representations when they appear. It also provides plotting, time windowing, and some other utility functions which are specifically intended for time series.
Relative to commercial packages the library is especially useful for time series research (such as studying estimation methods). For usual time series applications there may be commercial packages which are preferable.
The functions described in the Brief User's Guide should work fairly reliably, however, many of the functions described in the help facility are still under development and may not work. In addition, there may be functions described in the help facility for which the code is not yet distributed. This is a compromise which allows me to make the software available with minimum effort. This software is not a commercial product. It is the by-product of an ongoing research effort. Constructive suggestions and comments are welcomed. I can be reached at pgilbert@bank-banque-canada.ca or by phone at (613) 782-7346.
There is also a mailing list available which I hope will let users help
one another, as I cannot always be available. Please subscibe if you use
the library extensively, especially if you can help other users.
To subscribe send mail to
boc_list@bank-banque-canada.ca
with the body line subscribe boc\_dse Your Name
library("dse1") library("dse2")
The main objects are classes "TSdata", "TSmodel", and "TSestModel". These are, respectively, time series data objects, models, and objects with a model, data and some estimation information. For each of these there are several sub classes.
TSdata
,
TSmodel
,
TSestModel.object