Mean of classified data

I have a matrix X where each row represents an observation. For example:

X = 
     1 3
     2 5
     5 7

and a vector column classes that represents, for each observation in X, which class it belongs to, for example

classes=
     1
     2
     1

That means observation (1,3) and (5,7) are classified as 1 and observation (2,5) is classified as 2.

I'm looking for a efficiently method to calculate the mean of each class, so in this example, I get (3,5) mean for class 1 and (2,5) mean for class 2.

Answer 1

[Mann Baidi]

Hi Juanjo,

I understand you would like to perform “mean” operations on a classified 2D array.

For this you can use the “accumarray” function in MATLAB to perform arithmetic operations on classified arrays.

You can resolve your query using “accumarray” as follows:

X=[1 3;2 5; 5 7;10 15;64 67];
classes=[1 3 1 2 2]';
res=[accumarray(classes,X(:,1),[],@mean) accumarray(classes,X(:,2),[],@mean)]

res = 3×2

3 5 37 41 2 5

You can refer to the “accumarray” documentation using the following link.

https://www.mathworks.com/help/matlab/ref/accumarray.html

Hope this will help!