transform_signal.py

def transform_signal(signal, n_dim=160, min_max=(-1,1)):
    # convert data into -1 to 1
    signal_std = standardize_data(signal, min_data=min_num, max_data=max_num)
    # bucket or chunk size, 5000 in this case (800000 / 160)
    bucket_size = int(800000  / n_dim)
    # new_ts will be the container of the new data
    new_signal = []
    # this for iteract any chunk/bucket until reach the whole sample_size (800000)
    for i in range(0, 800000 , bucket_size):
        # cut each bucket to ts_range
        signal_range = signal_std[i:i + bucket_size]
        # calculate each feature
        mean = signal_range.mean()
        std = signal_range.std() # standard deviation
        std_top = mean + std # I have to test it more, but is is like a band
        std_bot = mean - std
        # I think that the percentiles are very important, it is like a distribuiton analysis from eath chunk
        percentil_calc = np.percentile(signal_range, [0, 1, 25, 50, 75, 99, 100]) 
        max_range = percentil_calc[-1] - percentil_calc[0] # this is the amplitude of the chunk
        relative_percentile = percentil_calc - mean # maybe it could heap to understand the asymmetry
        # now, we just add all the features to new_ts and convert it to np.array
        new_signal.append(np.concatenate([np.asarray([mean, std, std_top, std_bot, max_range]),percentil_calc, relative_percentile]))
    return np.asarray(new_signal)

ok