is there a way for a creation of variables for machine learning inside a loop from a dataset of waves

i have this dataset of wavelengths, problem is they are flipped and as such whenever i load the data i run a secondary part to rotate and flip, though so far i only have a way to read it and not really make use of it inside an ai machine learning thing

clc; clear; close all;

load('-mat','sub-036_task-eyesclosed_eeg.set','data','group','subject')

data_plot=flip(rot90(data,1));

N=width(data_plot);
B=height(data_plot);

figure 
%----------------------- all of this is just for plotting and seeing it in stackedplot
T = array2table(data_plot, 'VariableNames',"CH: "+(1:N));
s = stackedplot(T);
s.AxesProperties(1);
ax = findobj(s.NodeChildren, 'Type','Axes');
set(ax, 'YTick', []);
%--------------------------
xlim([-500 B+500])
grid on
set (zoom(gcf), 'Motion', 'horizontal', 'Enable', 'on');
%--------------------------

i was wondering if there is a way to make like a for loop so in each loop i get data_1 then data_2 data_3 etc etc etc or how could i use a machine learning or signal labeler app or ai or artificial neural network and deal with the issue that i need to rotateflip the data?

the database is from https://openneuro.org/datasets/ds004504/versions/1.0.7

[Stephen23]

"i was wondering if there is a way to make like a for loop so in each loop i get data_1 then data_2 data_3 etc etc etc "

The most important change is to always LOAD into an output variable:

S = load(..);

which you can then trivially loop over the fieldnames:

F = fieldnames(S);
for k = 1:numel(F)
    A = S.(F{k});
    .. do whatever with A
end

See also:

https://www.mathworks.com/help/matlab/ref/fieldnames.html

https://www.mathworks.com/help/matlab/matlab_prog/generate-field-names-from-variables.html

[Ri go]

this seems doable, will give it a try

Answer 1

[Walter Roberson]

data_names = who('-file', 'sub-036_task-eyesclosed_eeg.set', '-regexp', '^data_\d+$');
for DNI = 1:length(data_names)
    thisvar = data_names{DNI};
    DS = load('-mat','sub-036_task-eyesclosed_eeg.set', thisvar);
    data = DS.(thisvar);
    %stuff
end

[Ri go]

thank you for the reply, though this didn't seemed to work, what exactly it does?

[Walter Roberson]

The code certainly seems to work to me

SS.data_1 = [1 2 3];
SS.data_2 = [4 5 6];;

save('sub-036_task-eyesclosed_eeg.set', '-mat', '-struct', 'SS' );
whos -file sub-036_task-eyesclosed_eeg.set

  Name        Size            Bytes  Class     Attributes

  data_1      1x3                24  double              
  data_2      1x3                24  double              

data_names = who('-file', 'sub-036_task-eyesclosed_eeg.set', '-regexp', '^data_\d+$');
for DNI = 1:length(data_names)
    thisvar = data_names{DNI};
    DS = load('-mat','sub-036_task-eyesclosed_eeg.set', thisvar);
    data = DS.(thisvar);
    %stuff
    disp(data)
end

     1     2     3

     4     5     6

The code locates all of the data_ variables in the .set file. It loads them one by one into the variable data from where you can do <whatever> with them.

[Walter Roberson]

data_plot=flip(rot90(data,1));

I have to wonder whether that is correct code, or if instead the code should be

data_plot = flip(data.');

[Ri go]

thank you so much for the answers, sorry for not replying earlier, and well the rot90 and flip is there because of how the data is organized from the openneuro site, which basically had the x and y inverted

[Walter Roberson]

x = [1 2 3 4];
y = [100 200 300 400];
data = [x; y];
data

data = 2×4

1 2 3 4 100 200 300 400

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

flip(data)

ans = 2×4

100 200 300 400 1 2 3 4

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

data.'

ans = 4×2

1 100 2 200 3 300 4 400

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

flip(data.')

ans = 4×2

4 400 3 300 2 200 1 100

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

rot90(data,1)

ans = 4×2

4 400 3 300 2 200 1 100

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

flip(rot90(data,1))

ans = 4×2

1 100 2 200 3 300 4 400

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

so data.' gives the same result as flip(rot90(data,1)).

[Stephen23]

Note that LOAD accepts a regular expression directly:

data = load('sub-036_task-eyesclosed_eeg.set', '-regexp','^data_\d+$');

https://www.mathworks.com/help/matlab/ref/load.html#btm0etn-1-variables