## Find average y value of a range of numbers from a non linear data set

### math seeker (view profile)

on 20 Aug 2019
Latest activity Commented on by math seeker

on 21 Aug 2019

### the cyclist (view profile)

I have a data set of around 200000 lines. The data has 2 columns say x and y. I am giving a simplified example of the same below. X = [1, 5, 7, 12, 18, 25] Y= [20, 25, 32, 28, 27, 31]
This means for x values 1 to 4 y value is 20. For x =5 and 6 y = 25. For x=7 to 11 y =25 like that. Now I want to calcate average of subsets of x with 5 numbers. For example average of y of x values 1:5, then 6:10, 11:15 etc. For example, I expect average of 6:10 subset as [25*1(y value of x=6) + 32*4(y value of 7,8,9,10)] ÷ 5. What is the simplest way for doing this in matlab. I am a beginner in matlab programming.

### the cyclist (view profile)

on 20 Aug 2019
Edited by the cyclist

### the cyclist (view profile)

on 20 Aug 2019

Your example of averaging the values from 6:10 would be
mean(interp1(X,Y,6:10,"previous"))
Change where I put 6:10 to be the actual range you want.
The documentation for interp1 will explain the "previous" method of interpolation, which is crucial to how this works.
You can do multiple sets of 5 at once, like this:
mean(interp1(X,Y,[1:5; 6:10]',"previous"))
The first number is the average of the first 5 elements, and the second one is the next 5.
It's easy to generalize this even further.

math seeker

### math seeker (view profile)

on 21 Aug 2019
Thank you very much for such a simple answer

### MARTIN FABRET (view profile)

on 20 Aug 2019

function m = Meanfcn(X,Y,start,last)
firstIndex = 1;
while start < X(firstIndex)
firstIndex = firstIndex + 1;
end
l = last-start;
s = 0;
for i=0:1:l
if start+i>=X(firstIndex) && start+i<X(firstIndex+1)
s=s+Y(firstIndex);
else
firstIndex = firstIndex+1;
s=s+Y(firstIndex);
end
end
m = s/(l+1);
end