# Moving Average using a for loop

5 views (last 30 days)
Lisa Justin on 22 Apr 2014
Edited: Jan on 22 Apr 2014
Hi, How can I compute the moving average using a for loop and without using convolution (i.e conv()) or the filter function (i.e filter())?
Jan on 22 Apr 2014
If this is a homework question, and it does sound like it is one, please state this explicitly. Then show us, what you have tried so far and ask a specific question. We are not going to solve your homework.
Jan on 22 Apr 2014
Did you try anything to solve this problem by your own? What about asking your favorite internet search engine for Matlab code for a moving average? You'd find many many solutions in the FileExchange.

Star Strider on 22 Apr 2014
This seems to work:
data = rand(1,50); % Data
Dlen = length(data);
fltr = ones(1,5); % Filter
Flen = length(fltr);
datax = [data ones(1,Flen)*mean(data)]; % Pad ‘data’ vector with ones*mean(data)
mavg = [];
for k1 = 1:(Dlen)
mavg = [mavg (fltr * datax(k1:k1+Flen-1)')/Flen];
end
figure(1)
plot([1:Dlen], data)
hold on
plot([1:Dlen], mavg, '-r')
hold off
grid
Jan on 22 Apr 2014
Edited: Jan on 22 Apr 2014
The iterative growing of the output vector is extremely inefficient. Look for the term "pre-allocation" in this forum to find faster methods.
You can avoid a lot of divisions, if you apply the /FLen once to fltr instead of doing this in each iteration.