Incorrect use of the function zeros().

I have tried the following piece of script and got problems related to the use of zeros():

X = X(:);

N = length(X);

measurement_values(n)= zeros(1,tau); % tau=20

for n = 1:tau

y_tau=zeros(1,N/tau); % Line 27 ---> Error appears here.

for j = 1:N/n

y_tau(j) = mean(X(((j-1)*n+1):j*n));

end

measurement_values(n) = SampEn(y_tau,r,m);

end

Error report:

Error using zeros Size inputs must be integers.

Error in MSE (line 27) y_tau=zeros(1,N/tau);

Is there a way to fix the problem?

I checked the tutorial but still could not solve the problem.

Many thanks,

Massilon

Answer 1

[Stephen23]

As the error message clearly states, the inputs to zeros must be integer values (e.g. 0, 1, 2, ...). With your code you are supplying a fractional value of N/tau:

y_tau = zeros(1,N/tau)

which does not make any sense because how many zeros would you expect from, for example:

zeros(1,2.5)

???

The solution is to ensure that you supply integer values to zeros.

[Massilon Toniolo da Silva]

Hi Stephen, thanks for your answer. Yes, I agree completely. This code is not my own code. Surely, something is missing. I have thought to use the ceil(). But then there is a problem caused by too many arguments. Is there still a way of using the function ceil()? Many thanks, Massilon

[Stephen23]

y_tau = zeros(1,ceil(N/tau))