defining the mean of a values inside a matrix

Greetings,

I was trying to find a way to find the mean of the values inside the matrix. For example, [1,2,3,4,5,6] I want to find the average between 1 & 2/ 2&3/3&4/4&5 and so on

I tried to do a for-loop but It doesnt seemt to work when I compare the values to hand calculations

z18=(-length(theta)/2+(0:length(theta)))'*h; % Travel of total 18 layers %%%% 19 values due to neutral axis
z18t=((-length(theta)/2+(0:length(theta)))'*h)/2; %% I took this approach as well but it does not work 

for t=1:length(z18)
    tt=max(1,(t-1));
    ttt=min(length(z18), t+1);
    tttt(t)=mean(z18(tt:ttt));
end

[Walter Roberson]

movmean() ?

Answer 1

[Walter Roberson]

A = randi(9, [1 6])

A = 1×6

8 2 2 3 6 3

[A(1), A(1:end-1) + diff(A)/2, A(end)]

ans = 1×7

8.0000 5.0000 2.0000 2.5000 4.5000 4.5000 3.0000

[Eddy Ramirez]

this works as magic!

thank you

Answer 2

[Khalid Mahmood]

Further methods

1: movmean(matrix,windowsize) it calculates running mean but considers only windowsize elements

examples:

mm2=movmean(z18,2)
mm3=movmean(z18,3)

Using for loop to calculate running mean as [z18(1), (z18(1)+z18(2)/2, ( (z18(1)+z18(2) )/2+ z18(3) )/3,...]

mm=z18(1); 
for p=2:length(z18)
    mm(p)=(mm(p-1)+z18(p))/p;
end

[Sufia Tanveer]

Wow, here I find great programmers of Matlab... Great learning here. Awesome answers in quick time

[Sufia Tanveer]

I mean excellent answers

[Walter Roberson]

z18 = randi(9, [1 6])

z18 = 1×6

4 3 5 5 7 7

mm=z18(1); 
for p=2:length(z18)
    mm(p)=(mm(p-1)+z18(p))/p;
end
mm

mm = 1×6

4.0000 3.5000 2.8333 1.9583 1.7917 1.4653

movmean(z18,2)

ans = 1×6

4.0000 3.5000 4.0000 5.0000 6.0000 7.0000

[z18(1), z18(1:end-1) + diff(z18)/2, z18(end)]

ans = 1×7

4.0000 3.5000 4.0000 5.0000 6.0000 7.0000 7.0000

I don't think the mm code is correct at all.

[Sufia Tanveer]

At 1st I also thought that mm code seems wrong. But it is desired sometimes to reduce values in this manner. I like that way of reducing values... It is not a mean but a way to reduce values

[Khalid Mahmood]

Yes, both of you are correct. Above mm code uses previous average and next value of vector, which off course reduces values. A better way to use all previous values of vector (not previous average) in moving mean is :

z18 = randi(9, [1 6]);
mm=z18(1); 
for p=2:length(z18)
    mm(p)=sum(z18(1:p))/p;
end
z18

1 3 5 9 9 2

mm

1 2 3 4.5 5.4 4.8333

and vectorised form is

mm=cumsum(z18)./(1:length(z18))

1 2 3 4.5 5.4 4.8333

[Sufia Tanveer]

Its great now!

[Sufia Tanveer]

I learnt more from comments than actual question!

[Khalid Mahmood]

And remember, In moving mean, last value of source vector is not copied in result. It is actually calculated. So moving mean methods, movmean, mm code (to rduce in a mean way), and vectorised cumsum./[1:src] methods are actual moving mean methods. Those are just different types of moving means