ARFIMA(p,d,q) estimator

Maximum likelihood estimators of stationary univariate ARFIMA(p,d,q) processes.

Description

ARFIMA(p,d,q) maximum likelihood estimators:

- Whittle estimator
- Exact maximum likelihood estimator

-and some other,possibly useful functions,forecasting included.I've yet to implement the prediction error bands calculation (as one can judge from the screenshot).

Requirements:

-Statistics Toolbox
-Optimization Toolbox
-Kevin Sheppard's MFE Toolbox

http://www.kevinsheppard.com/wiki/MFE_Toolbox

Optional requirements:

-Simone Fatichi's ARFIMA(p,d,q) simulator(MATLAB Central FileExchange #25611)

The latter is required for testing the algorithms' performance (implemented in arfima_test).

NOTICE: this time, a C/MEX file is there to speed up the process,without a .m equivalent.It compiled with several compilers (LCC and MS VC++ 2008) and proved to be stable for me.

Further updates are planned:

-other estimation algorithms
-documentation
-etc.

Any comments,suggestions to:

[e-mail]

Acknowledgements

The author wishes to acknowledge the following in the creation of this submission:
MexWizard, howto: mex + MS Visual C/C++ Express Edition 2005, C++ MEX symbolic debugging, ARFIMA simulations

• Required Products: Statistics Toolbox
• MATLAB release: MATLAB 7.6 (R2008a)

[Evan]

How do I use these parameters to actually make a prediction?

[György Inzelt]

Hi Evan,added a forecasting function.Hope you'll find it helpful.

[name1 name2]

I don't know how to compile C++ file with matlab can someone help me please?
herés my mail [e-mail]