How can I create a vector of rolling Standard Deviations

Hi !

I'am sure this is an easy question for most of you...

I have a timeseries 'T' and I wanted to convert this into a vector 'V' of rolling Standard deviations. So every element of 'V' is the Std of the previous 30 elements of 'T'. The first thirty elements of 'V' are NaN then, right?

Could someone pleas help me with this problem?

Thanks, Nico

Answer 1

[Shashank Prasanna]

It is preferred if you give it a shot first and then seek help,

w = 30;
y = NaN(100,1)
x = randn(100,1);
for ii = w+1:100
    y(ii,1) = var(x(ii-w:ii-1,1));
end

The first 30 need not be NaN, you can make them zero if you prefer.

[Nico]

Thank you - that helps !