***************************************************************

  Example Codes for Explicit Link Between Periodic Covariance 
              Functions and State Space Models

***************************************************************

INTRODUCTION

  This package acts as a supplementary code package for the 
  following paper:

  @inproceedings{Solin+Sarkka:2014,
     author = {Solin, Arno and S\"arkk\"a, Simo},
      title = {Explicit link between periodic covariance 
               functions and state space models},
  booktitle = {Proceedings of the 17th International Conference
               on Artificial Intelligence and Statistics},
     series = {{JMLR} Workshop and Conference Proceedings},
       year = {2014},
     volume = {33}
  }

  Available online at: 
    http://jmlr.org/proceedings/papers/v33/solin14.html

  If you find the methods useful in your research or work, 
  please cite the paper.


DESCRIPTION:

  This code package features the methods for converting periodic
  covariance functions to state space models. These codes are
  compact and available as two separate files: `cf_periodc_to_ss.m'
  and `cf_quasiperiodic_to_ss.m'.

  In order to demonstrate the use of the models in two real-world 
  settings, we have additionally added codes for doing the actual 
  inference and forming the models. This requires us to include a 
  lot of additional 'house-keeping' code as well.


PACKAGE CONTENTS:

* CORE

  Code for converting covariance functions in Gaussian process
  regression to state space models ('cf_*_to_ss.m') and codes
  for doing inference using the state space models ('ss_*.m').

* MAUNALOA

  Codes for the Mauna Loa CO2 example in the paper. See the
  code and the publication for details.

* BIRTHS

  Codes for the birth date example in the article. See the
  code and the publication for details.


NOTES:

* The codes were tested under Matlab 2012a they should to
  most part be compatible with Octave as well, but the 
  behavior of some special functions (e.g. `lyap' differs).

* These methods are also included as a part of the GPstuff 
  toolbox now. For doing inference in practice, that 
  implementation is perhaps more user-friendly.

* This package features some preliminary code from the SGP
  (sequential GP) toolbox that is scheduled to be published
  later on in 2014. Details on the inference algorithms
  will be found in the SGP documentation. Contact the 
  authors for more details.
  

VERSION:

  This code was originally uploaded: 2014-04-16

COPYRIGHT:
  
  (c) Arno Solin and Simo Särkkä, 2014

      The codes from the SGP toolbox are copyrighted by 
      Arno Solin, Jukka Koskenranta and Simo Särkkä.

      The data and the GPstuff functions are copyrighted 
      by their respective copyright holders.


LICENSE:

  This software is distributed under the GNU General Public
  License (version 3 or later); please refer to the file
  license.txt, included with the software, for details.  



