Concatenated Segmentation using Overlapping Windows

I have a N-dimensional timeseries data with m samples (Nxm). I would like to sample the table with a fixed sized window of 50 samples with a shift of 5 samples, concatanating each sample in another table. What is an efficient algorithm to run this process?

[KALYAN ACHARJYA]

Can you explain with numeric data, considering any sample 1D array data?

Answer 1

[Mathieu NOE]

hello

example below :

% dummy data
data = rand(80,5);
shift = 5;    % nb of samples for averaging 
buffer = 50;    % nb of samples for averaging 

for ci=1:fix((length(data)-buffer)/shift +1)
    start_index = 1+(ci-1)*shift;
    stop_index = min(start_index+ buffer,length(data));
    out_data{ci} =data(start_index:stop_index,:);  % 
    figure(ci),
    plot(out_data{ci});
end