# moving average considering the first and end data points without changing window size, cyclic window

6 views (last 30 days)
nlm on 20 Mar 2020
Commented: Rik on 21 Mar 2020
How can you calculate moving average, where for edges, it uses the first and last data (cyclic) points instead of decreasing the window size. For example, if my window size = 10, and number of data points = 100, then for an moving average at point 1, it uses end 5 values at 100,99,98,97,96 and 2,3,4,5,6. Similarly, at point 2, it uses, 1,100,99,98,97 and 3,4,5,6,7 so on and so forth.

Ameer Hamza on 20 Mar 2020
Edited: Ameer Hamza on 21 Mar 2020
Try this
x = 1:100;
window_size = 11;
window = ones(1, window_size)/window_size;
y = cconv(x, window, numel(x));
y = circshift(y, -floor(window_size/2));
Without cconv
x = 1:100;
window_size = 11;
left_window = floor((window_size-1)/2);
right_window = floor((window_size)/2);
x = [x(end-left_window+1:end) x x(1:right_window)];
y = movmean(x, [left_window right_window], 'Endpoints', 'discard');
Rik on 21 Mar 2020
I didn't know about the cconv function, so my first thought was to extend both side with the data from the other side (so your second option). I you keep learning