I'm getting this as error on 2nd line --> Index in position 2 exceeds array bounds (must not exceed 1).

ecg_nsr = load('ecg_nsr.mat')
mat=ecg_nsr(:,3); % mat=ecg_nsr(:,1:15);

[r,c]=size(mat);
%r=1800;

wsize=50;
st=1:wsize:r;
en=wsize:wsize:r;

if(length(st)>length(en))
    en=[en,r];
end

% if((st(end)-en(end))<c)
%     st(end)=[];
%     en(end)=[];
% end

result={};
F={};
Fmat={};

len=length(st);
reg=reg_mat;
parfor i=1:len
    i
    ori_mat=mat(st(i):en(i),:);
    reg=reg_mat(st(i):en(i),:);
    [FitArray,FitMat, offsprings]=permute_one(ori_mat,mat, reg);
    
    F{i}=FitArray;
    result{i}=offsprings;
    Fmat{i}=FitMat;
end

Answer 1

[DGM]

The functional syntax for load behaves differently than the command syntax. Say you have a matfile containing arrays X,Y, and Z. If using the command syntax as follows

load myfile.mat

The arrays X, Y, Z will just magically appear in the workspace. This is both attractive in its simplicity and incredibly problematic. Just as I can't look at the code and know what variables are (or aren't) going to show up, neither can MATLAB. This is discouraged practice; kudos for not doing it!

On the other hand, using the functional syntax like this

S = load('mymatfile.mat');

will load X,Y, and Z into a the scalar struct S. They could be accessed as S.X, S.Y, and S.Z.

So now we have this:

ecg_nsr = load('ecg_nsr.mat')
mat = ecg_nsr(:,3);

ecg_nsr is a scalar struct, so it doesn't have a third column to index into. If the contents of the matfile is a single array called "ecg_nsr", then you can use it in-place, or make a copy.

S = load('ecg_nsr.mat')
mat = S.ecg_nsr(:,3); % use in-place