Please expalin the error

Hi ,

for the data set y, I need to fit an ar model and calculate the best order using aic.

I used the following code to do that:

V = arxstruc(y,struc(1:15));
m = selstruc(V,'aic');

However, I got the following error when I tried to run it.

Error using struc (line 21)
Not enough input arguments.

Error in big100 (line 47)
V = arxstruc(y,struc(1:15));

Can someone correct this please?

Full code:

clc;
clear;

a0 = 0.05; a1 = 0.1; b1 = 0.85;

epsi=zeros(3000,1);
simsig=zeros(3000,1);

for i = 1:3000
                 if (i==1) 
                     simsig(i) = a0 ;
                     s=(simsig(i))^0.5;
                     epsi(i) = normrnd(0,1) * s;
                 else
                     simsig(i) = a0+ a1*(epsi(i-1))^2+ b1*simsig(i-1);
                     s=(simsig(i))^0.5;
                     epsi(i) = normrnd(0,1) * s;
                 end
end

yt1=epsi.^2;

y = yt1(2001:3000);

V = arxstruc(y,struc(1:15));
m = selstruc(V,'aic');

Answer 1

[Brian B]

The documentation for struc indicates that it takes three input arguments:

 >>help struc

 struc  Generate typical structure matrices for ARXSTRUC and IVSTRUC.
    NN = struc(NA,NB,NK)
    NA, NB and NK are vectors containing the orders na, nb and delays nk
    to be tested. See help on ARX for an explanation of na, nb and nk.
    NN is returned as a matrix containing all possible combinations of these
    orders and delays.

    ...

But you have called it with only one vector (1:15). You need to choose values for nb and nk to test as well.

[dav]

even if I put the same range for nb and nk I still get the error..

[Brian B]

The same error? If I enter

 struc(1:15, 1:15, 1:15)

I get a matrix whose rows are of the form (na, nb, nk). It is not surprising if you still get an error, though, since your calling syntax for arxstruc does not look right either.

Type

 help arxstruc

to see how to correct that.

[Brian B]

There are two datasets (denoted ZE and ZV in the documentation) and one matrix of orders (denoted NN). The matrix of orders is what you get from struc. The two datasets are used for estimation (ZE) and validation (ZV). Note that these are not just vectors, they are IDDATA objects.

Once more:

 help iddata

If you want to use the same data for both, pass it in twice.