How do I find the mean of certain parts of a cell array?

Hello. I am trying to find the mean value of certain parts of a cell array. Each cell array row has 8192 values and I am trying to find the mean of each.

A snippet of the cell array:

For example the values of the first 2 rows of the cell array are:

I want to find the mean value of each column from a range of cell array rows 3-10.

Any ideas on how to do so?

Thank you very much!

Answer 1

[Jan]

This is easy, if you store the values in a numerical array instead of a cell. Why do you prefer a cell array?

But you can create the numerical matrix dynamically (although this is not really fast):

% mean value of each column from a range of cell array rows 3-10:
X = mean(cat(1, YourCell{3:10}), 1);

[Michael Arvin]

I chose cell array because I took the suggestion from the answer in this https://www.mathworks.com/matlabcentral/answers/1657370-how-do-i-input-multiple-values-in-an-array-matrix-multiple-times?s_tid=srchtitle

If you do have a suggestion on how to do it better, I would greatly appreciate it.

[Jan]

If all cell arrays contain vectors of the same size, using a matrix is more efficient.

cat(1, YourCell{3:10}) creates such a matrix, but if you want to apply more operations, convert the full cell array instead:

X = cat(1, YourCell{:});

mX = mean(X(3:10, :), 1)

[Michael Arvin]

Got it! Thank you.

[Jan]

You are welcome. Finding an efficient representation for the data is a crucial task in programming. Cell arrays are useful, if they contain elements of different sizes or classes. If all have the same size and class, a multi-dimensional array is (usually) better - except if it is huge and a set of smaller arrays would match into the RAM more likely.

Do I use an array of structs or a struct of arrays? Should I use the solution, which uses less memory or is faster to process? On one hand efficient code helps to save minutes of runtime, on the other hand a clean and clear code can avoid hours of debugging. Such questions belong to the work of an experiences programmer. And sometimes the best choice is not to decide for a specific representation, but to keep it flexible for later changes.