python 计量做hp滤波_MATLAB解决HP滤波 急啊

function [x]= penta2(y,a,b,c)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Author: Kurt Annen annen@web-reg.de

% Date: 15/05/2004

% Internet: www.web-reg.de

%

% Solves the problem Ax=b when A is pentadiagonal and strongly nonsingular.

% This is much faster than x=A\y for large matrices.

%

% Reference: Sp䴨, Helmuth "Numerik: Eine Einf? f?hematiker und Informatiker"

%               S. 110 . Vieweg-Verlag Braunschweig/Wiesbaden (1994)

%

% a = main diagonal

% b = 2. diagonal

% c = 3. diagonal

%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

if nargin ~= 4

error('penta(A,y) requires four arguments');

end

[n] = length(a);

[m] = length(y);

[o] = length(b);

[p] = length(c);

if m ~= n

error('Inner matrix dimensions must agree');

end

c(m)=0;

c(m-1)=0;

b(m)=0;

if m ~= length(b) &  m ~= length(c)

error('a,b,c must have the same dimension');

end

% a optimized algorithm

h1=0;

h2=0;

h3=0;

h4=0;

h5=0;

hh1=0;

hh2=0;

hh3=0;

hh4=0;

hh5=0;

z=0;

hb=0;

hc=0;

for i=1:m

z=a(i)-h4*h1-hh5*hh2;

hb=b(i);

hh1=h1;

h1=(hb-h4*h2)/z;

b(i)=h1;

hc=c(i);

hh2=h2;

h2=hc/z;

c(i)=h2;

a(i)=(y(i)-hh3*hh5-h3*h4)/z;

hh3=h3;

h3=a(i);

h4=hb-h5*hh1;

hh5=h5;

h5=hc;

end

h2=0;

h1=a(m);

y(m)=h1;

for i=m:-1:1

y(i)=a(i)-b(i)*h1-c(i)*h2;

h2=h1;

h1=y(i);

end

x=y;

end

return