Description of the program: arima-model

This program fits (by means of least squares) an autoregressive integrated moving average (ARIMA) model to the possibly multivariate data.
As a first step a AR model is fitted to give a first guess of the residuals which enter the MA part. With these residuals the full ARIMA is fitted. This fit is repeated until convergence of the residuals is reached or a maximum number of iterations was performed.
Note that no attempt has been made to generate a stable model.

Usage:

arima-model [Options]

Everything not being a valid option will be interpreted as a potential datafile name. Given no datafile at all, means read stdin. Also - means stdin

Possible options are:

Option Description Default

-l# number of data to use whole file

-x# number of lines to be ignored 0

-m# dimension of the vectors 1

-c# column to be read 1,...,dimension of the vectors

-p# order of the initial AR-model 10

-P# order of the AR,I,MA model 0,0,0 (means it just does the initial AR Modeling)

-I# max. number of iterations of the ARIMA Fit 50

-e# required accuracy of the ARIMA convergence 0.001

-s# length of iterated data set no iteration

-o# output file name without file name: 'datafile'.ari
(or stdin.ari if stdin was used)
if no -o is given stdout is used

-V# verbosity level
  0: only panic messages
  1: add input/output messages
  2: print residuals though iterating a model
  4: print original data + residuals 1

-h show these options none

Option	Description	Default
-l#	number of data to use	whole file
-x#	number of lines to be ignored	0
-m#	dimension of the vectors	1
-c#	column to be read	1,...,dimension of the vectors
-p#	order of the initial AR-model	10
-P#	order of the AR,I,MA model	0,0,0 (means it just does the initial AR Modeling)
-I#	max. number of iterations of the ARIMA Fit	50
-e#	required accuracy of the ARIMA convergence	0.001
-s#	length of iterated data set	no iteration
-o#	output file name	without file name: 'datafile'.ari (or stdin.ari if stdin was used) if no -o is given stdout is used
-V#	verbosity level 0: only panic messages 1: add input/output messages 2: print residuals though iterating a model 4: print original data + residuals	1
-h	show these options	none

Description of the Output:

If P is not given:
The first line contains the forecast error averaged over all components. The second line contains the average forecast errors for all components individually. The third line contains the Log-likelihood and the AIC values of the fit. The next p*m lines contain the fitted ar coefficients and the rest of the file are either the residuals of the fit or a simulated new trajectory.

If P is given:
The first lines (marked #iteration xxx) show the convergence of the residuals of the ARIMA fit. The rest of the files is like in the above case.

View the C-sources

Table of Contents * TISEAN home