Adding Arrays of cells

I have ten arrays of 15-dimensional A1, ..., A10 which consist of cells of different size. All arrays have same format i.e., same dimension of cells for all Ai, i=1,..10 respectively.

Ai = [2x1 double] [1x2 double] [2x2 double] [2x1 double] [2x2 double] [2x2 double] [2x2x2 double] [2x1 double] [2x2 double] [2x2 double] [2x2x2 double] [2x2 double] [2x2x2 double] [2x2x2 double] [4-D double]

I want A in same format where A is an array which have the mean of all components of cells of Ai, i =1,...,10. That is, A = (A1+ ... +A10)/10.

Please help me. Thanks :MSA

[Stephen23]

"I have ten arrays of 15-dimensional A1, ..., A10..."

And that is the start of your problem right there. Do not create lots of numbered variables! Simply store your data in one array and your life will be much simpler (and your code faster and less buggy too).

This is what the MATLAB documentation says about your code idea: "A frequent use of the eval function is to create sets of variables such as A1, A2, ..., An, but this approach does not use the array processing power of MATLAB and is not recommended. The preferred method is to store related data in a single array."

source: https://www.mathworks.com/help/matlab/matlab_prog/string-evaluation.html

You can read more explanations of the problems caused by accessing variable names dynamically:

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

Answer 1

[Guillaume]

There is a reason we keep on saying not to create numbered or dynamically named variables. It makes using them extremely difficult.

First step: change your algorithm to create just one variable. Options are a cell array of these cell arrays or concatenating the content of the cells in an extra dimension. Once that's done it's almost trivial to calculate the mean:

A = {A1, A2, ...}  %cell array of cell arrays
meanA = cellfun(@(varargin) mean(cat(ndims(varargin{1})+1, varargin{:}), ndims(varargin{1})+1) , A{:}, 'UniformOutput', false)

The complication here is that I don't know the dimensionality of the matrices you want to average. All the cellfun does is concatenate them all in a higher dimension and take the mean in that dimension.

The result of the concatenation without the mean calculation would be the other way to keep your matrices together

A = {A1, A2, ...}  %cell array of cell arrays
A = cellfun(@(varargin) cat(ndims(varargin{1})+1, varargin{:}), A{:}, 'UniformOutput', false)