How do I take the average of elements of a vector depending on the values of another vector?

Given two vectors, how do I take the average of values in one vector depending on what the value is in the second vector?

For example:

A = [2 3 2 2 3 4 3 3];

B = [0.1 0.2 0.22 0.13 0.07 0.88 0.3 0.5];

Here, I'd want to take the average of elements 1, 3, and 4 in B because they correspond to A = 2, and then similarly elements 2, 5, 7, and 8 in B that correspond to A = 3, and then element 6 corresponding tto A = 4.

Answer 1

[Torsten]

A = [2 3 2 2 3 4 3 3];
B = [0.1 0.2 0.22 0.13 0.07 0.88 0.3 0.5];
Au = unique(A,'stable');
C = arrayfun(@(i)mean(B(A==Au(i))),1:numel(Au))

C = 1×3

0.1500 0.2675 0.8800

[Torsten]

I modified the answer from

Au = unique(A);

to

Au = unique(A,'stable');

for that the mean values are sorted according to the values appearing in A.

I don't know if this is important for your application.

Answer 2

[Paul]

This common workflow is discussed here: splitapply

A = [2 3 2 2 3 4 3 3];
B = [0.1 0.2 0.22 0.13 0.07 0.88 0.3 0.5];
[G,ID]=findgroups(A);
groupmean = splitapply(@mean,B,G);
table(ID.',groupmean.','VariableNames',{'ID' , 'GroupMean'})

ans = 3×2 table
    ID    GroupMean
    __    _________

    2        0.15  
    3      0.2675  
    4        0.88  

[Torsten]

It might be necessary to order the groups according to the occurence of the elements in A.

Thus

findgroups([4 2 3 2 2 3 7 5 3 3])

should produce

[1 2 3 2 2 3 4 5 3 3 ] 

not

[3 1 2 1 1 2 5 4 2 2]