Here’s how you’d implement a low pass butterworth filter in matlab. This has come in very handy.
lowcutoff=1000; sampling_freq=12000; fNorm=lowcutoff/(sampling_freq/2); [b,a] = butter(10, fNorm, 'low'); filtered_signal = filtfilt(b, a, signal);
Here’s how you’d add (gaussian white) noise:
Matlab can be a real pain when trying to save CSV files with different data types (string + numeric: lets say). But to try things like SVN and Naive Bayes on weka, you need your labels to be nominal (not numeric). Fortunately, this guy wrote this code which works like a charm. All you have to do (after downloading that file) is :
cell2csv ( 'filename.csv', cell_array_that_you_want_to_write, ' , ');