sw_d.m

计量工具箱

% PURPOSE: demo of sw()
%         Temporal disaggregation using the Stram-Wei method

%---------------------------------------------------
% USAGE: sw_d
%---------------------------------------------------

close all; clear all; clc;

% Low-frequency data: Spain's Exports of Goods. 1995 prices

Y=[  20499
     23477
     25058
     27708
     31584
     31898
     30233
     32235
     34049
     36035
     39795
     44299
     47426
     52339
     62949
     69885
     77174
     90133
     96496
    102776
    113026
    115573 ];

% ---------------------------------------------
% Inputs for td library

% Type of aggregation
ta=1;

% Minimizing the volatility of d-differenced series
d=1;

% Frequency conversion
s=4;

% Number of observations of low-frequency input
N = length(Y);

% Number of observations of high-frequency output
n = s*N;

% Defining the VCV matrix of stationary high-frequency time series
% Assumption of the example: IMA(d,2). For a comprehensive and more general analysis
% please consult Stram and Wei (1986)"Temporal aggregation in the ARIMA process",
% Journal of Time Series Analysis, vol. 7, n鷐. 4, p. 279-292.

th1 =  0.9552;
th2 = -0.0015;
va = 0.87242 * ((223.5965)^2);

acf0 = va * (1+th1^2+th2^2);
acf1 = -va * th1 * (1-th2);
acf2 = -va * th2;

a0(1:n-d)=acf0;
a1(1:n-d-1)=acf1;
a2(1:n-d-2)=acf2;

v=0.5*diag(a0)+diag(a1,-1)+diag(a2,-2);
v=v+tril(v)';

% Name of ASCII file for output
file_sal='sw.sal';

% Calling the function: output is loaded in a structure called res
res = sw(Y,ta,d,s,v);

% Calling printing function
tduni_print(res,file_sal);
edit sw.sal;

% Calling graph function
tduni_plot(res);