matlab - Efficient way to apply arrayfun to a matrix (i.e. R^N to R^M) -

i have function transforms r^n r^m. simplicity, lets let identity function @(z) z z may vector. want apply function list of parameters of size k x n , have map k x m output.

here attempt:

function out_matrix = array_fun_matrix(f, vals)     i=1:size(vals,1)         f_val = f(vals(i,:));         if(size(f_val,1) > 1) %need stack rows, convert required.             f_val = f_val';         end         out_matrix(i,:) = f_val;     end end 

you can try with

array_fun_matrix(@(z) z(1)^2 + z(2)^2 + z(3), [0 1 0; 1 1 1; 1 2 1; 2 2 2]) 

the question: there better , more efficient way vectorization, etc.? did miss built-in function?

examples of non-vectorizable functions: there many, involving elaborate sub-steps , numerical solutions. trivial example looking numerical solution equation, in term using numerical quadrature. i.e. let params = [b c] , solve a such int_0^a ((z + b)^2) dz = c

(i know here calculus, integral here stripped down). implementing example,

find_equilibrium = @(param) fzero(@(a) integral(@(x) (x + param(1)).^2 - param(2), 0, a), 1) array_fun_matrix(find_equilibrium, [0 1; 0 .8]) 

you can use cellfun function, you'll need manipulate data bit:

function out_matrix = array_fun_matrix(f, vals)     % convert data cell array:     cellvals = mat2cell(vals, ones(1,size(vals,1)));     % apply function:     out_cellarray = cellfun(f, cellvals, 'uniformoutput', false);     % convert matrix:     out_matrix = cell2mat(out_cellarray); end 

if don't implementation, can improve performance of yours preallocating out_matrix:

function out_matrix = array_fun_matrix(f, vals)     firstoutput = f(vals(1,:));     out_matrix = zeros(size(vals,1), length(firstoutput)); % preallocate speed.     i=1:size(vals,1)         f_val = f(vals(i,:));         if(size(f_val,1) > 1) %need stack rows, convert required.             f_val = f_val';         end         out_matrix(i,:) = f_val;     end end