1 ## Copyright (C) 1995-2012 Friedrich Leisch
3 ## This file is part of Octave.
5 ## Octave is free software; you can redistribute it and/or modify it
6 ## under the terms of the GNU General Public License as published by
7 ## the Free Software Foundation; either version 3 of the License, or (at
8 ## your option) any later version.
10 ## Octave is distributed in the hope that it will be useful, but
11 ## WITHOUT ANY WARRANTY; without even the implied warranty of
12 ## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 ## General Public License for more details.
15 ## You should have received a copy of the GNU General Public License
16 ## along with Octave; see the file COPYING. If not, see
17 ## <http://www.gnu.org/licenses/>.
20 ## @deftypefn {Function File} {[@var{d}, @var{dd}] =} diffpara (@var{x}, @var{a}, @var{b})
21 ## Return the estimator @var{d} for the differencing parameter of an
22 ## integrated time series.
24 ## The frequencies from @math{[2*pi*a/t, 2*pi*b/T]} are used for the
25 ## estimation. If @var{b} is omitted, the interval
26 ## @math{[2*pi/T, 2*pi*a/T]} is used. If both @var{b} and @var{a} are
27 ## omitted then @math{a = 0.5 * sqrt (T)} and @math{b = 1.5 * sqrt (T)}
28 ## is used, where @math{T} is the sample size. If @var{x} is a matrix,
29 ## the differencing parameter of each column is estimated.
31 ## The estimators for all frequencies in the intervals
32 ## described above is returned in @var{dd}. The value of @var{d} is
33 ## simply the mean of @var{dd}.
35 ## Reference: P.J. Brockwell & R.A. Davis. @cite{Time Series:
36 ## Theory and Methods}. Springer 1987.
39 ## Author: FL <Friedrich.Leisch@ci.tuwien.ac.at>
40 ## Description: Estimate the fractional differencing parameter
42 function [d, dd] = diffpara (x, a, b)
44 if ((nargin < 1) || (nargin > 3))
50 x = reshape (x, n, 1);
63 if (! (isscalar (a) && isscalar (b)))
64 error ("diffpara: A and B must be scalars");
67 dd = zeros (b - a + 1, k);
71 w = 2 * pi * (1 : n-1) / n;
73 x = 2 * log (abs (1 - exp (-i*w)));
74 y = log (periodogram (x(2:n,l)));
80 dd(m-a+1) = - x(1:m) * y(1:m) / sumsq (x(1:m));