same as solve_lsq_book_constsize, but G does now have varibale size overall we get N = numel(sizes) many least squares problem portions of sizes(i) of each input belong to the same problem this function returns the solution vectors and the condition numbers one can also set flags and parameters for regularization and similiar things
Syntax
[c_book, conds] = solve_lsq_book_varsize(weights, basis_values, f_values, sizes, varargin)