nested for loop for moving average filter

Hello Everyone,

I want to implement nested for loop in MATLAB the first loop should run from 3 to n and inside this loop the second for loop should run for 1:5. My problem is I want to take averages of 5 numbers but a moving average. So i should start with average from elements 1 to 5 for first row, elements 2 to 6 for second row , elemets 3 to 7 for third row.

Please help me with this problem.

Thanks in Advance

[Jan]

What have you tried so far? The text description seems to be clear enough to convert this directly to Matlab.

[Jatin Arora]

The problem was the length of initial array is not fixed and also the length of number of neighbouring points is not fixed. So I was not able to put a constant value. I was able to solve the problem

    for i = 3:end

              j1=i-No;     
              j2=i+No;

    for j = j1:j2
        PtsSum(i,:)=PtsSum(i,:)+PathPntIn(j,:); 
     end 
end

Now I am stuck at a different thing. So will put a new question.

[Image Analyst]

You always have to decide what to do at the end when the window falls outside the signal. Looks like you want to just limit the window so that it never goes outside your signal. I think you need to do

for i = 2 : end-2  % not 3:end

where No is 2, which is the half width of a 5 element wide window.

Answer 1

[Jan]

No loops required in Matlab:

n  = length(x);
mx = (x(1:n-4) + x(2:n-3) + x(3:n-2) + x(4:n-1) + x(5:n)) / 5;

[Daniel Shub]

I thought this would be faster than conv, but I was wrong. I wonder what the fastest method is for a moving average filter.

Answer 2

[Daniel Shub]

While you can do this with a loop, MATLAB has some highly optimized commands for filtering. The simpliest is probably conv.

First create some dummy date

N = 20;
x = randi(10, N, 1);

The M point moving average filter is

M = 5;
y = conv(x, ones(M, 1));