skdh.features.SpectralEntropy#

class skdh.features.SpectralEntropy(padlevel=2, low_cutoff=0.0, high_cutoff=5.0)#

A measure of the information contained in the power spectral density estimate. Similar to SignalEntropy but for the power spectral density.

Parameters:

padlevelint, optional: Padding (factors of 2) to use in the FFT computation. Default is 2.
low_cutofffloat, optional: Low value of the frequency range to look in. Default is 0.0 Hz
high_cutofffloat, optional: High value of the frequency range to look in. Default is 5.0 Hz

Attributes:

Methods

compute(signal[, fs, axis])

Compute the spectral entropy

Notes

The padlevel parameter effects the number of points to be used in the FFT computation by factors of 2. The computation of number of points is per

\[nfft = 2^{ceil(log_2(N)) + padlevel}\]

So padlevel=2 would mean that for a signal with length 150, the number of points used in the FFT would go from 256 to 1024.

compute(signal, fs=1.0, *, axis=-1)#

Compute the spectral entropy

Parameters:

signalarray-like: Array-like containing values to compute the spectral entropy for.
fsfloat, optional: Sampling frequency in Hz. If not provided, default is assumed to be 1Hz.
axisint, optional: Axis along which the signal entropy will be computed. Ignored if signal is a pandas.DataFrame. Default is last (-1).

Returns: