Spectral Features¶

best.feature_extraction.SpectralFeatures.mean_bands(args)¶

Mean power spectral density - Mean spectral power for each frequency band.

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.mean_frequency(args)¶

Mean dominant frequency - Calculates mean dominant frequency on a frequency range defined as min-to-max of frequency bands at the input. - Source: https://www.mathworks.com/help/signal/ref/meanfreq.html

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.median_frequency(args)¶

Spectral median frequency

Calculates median dominant frequency on a frequency range defined as min-to-max of frequency bands at the input.
Source: https://www.mathworks.com/help/signal/ref/medfreq.html

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.non_normalized_entropy(args)¶

Spectral entropy (Shannon Entropy)

Estimates Shannon Entropy of a spectrum on a frequency range defined as min-to-max of frequency bands at the input.
Source: https://www.mathworks.com/help/wavelet/ref/wentropy.html

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.non_normalized_entropy_bands(args)¶

Spectral entropy (Shannon Entropy)

Estimates Shannon Entropy of a spectrum on a frequency range defined as min-to-max of frequency bands at the input.
Source: https://www.mathworks.com/help/wavelet/ref/wentropy.html

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.normalized_entropy(args)¶

Spectral entropy (Shannon Entropy)

Estimates Shannon Entropy of a spectrum on a frequency range defined as min-to-max of frequency bands at the input.
Source: https://www.mathworks.com/help/wavelet/ref/wentropy.html

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.normalized_entropy_bands(args)¶

Spectral entropy (Shannon Entropy)

Estimates Shannon Entropy of a spectrum on a frequency range defined as min-to-max of frequency bands at the input.
Source: https://www.mathworks.com/help/wavelet/ref/wentropy.html

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names

best.feature_extraction.SpectralFeatures.relative_bands(args)¶

Relative spectral density

Mean spectral power for each frequency band relative to the power of a whole spectrum defined on a frequency range defined as min-to-max of frequency bands at the input.

Parameters

args (dictionary) –

‘psd’ (numpy.ndarray[n_samples, n_freq_samples]) - one-sided PSD
’fbands’ (list of lists) - frequency bands in which the feature is to be calculated [[0.5, 4], [5, 9]]
’freq’ (numpy.array[n_freq_samples]) - reference frequency array for the PSD

Returns

x (list(numpy.array)) – Calculated features for individual frequency bands
feature_name (list(numpy.array)) – Feature names