eva_classes module

class eva_classes.Processing(signals, timestamps)

Bases: object

The class for processing the Signals. It takes the signal as an array and the timestamps as a pandas.Series objekt. Included are:

pleth_filter

resampling

slice_list

pleth_filter(fs, lowcut, highcut, order, resampled=False)

Butterwoth IIR filter. It takes the signal and filters it in a second order section mannor. If the signal was resampled with Processing.resampled, resampled must be set True. The filtered signal is stored in self.filtered_signal.

Parameters

fs_AInt

Sampling rate in Hertz. The samplingrate is provided with the config.ini. For more information on that go to PPG_EVA_GUI.set_values().

lowcutfloat, optional

The lower bound of the Filter. It is provided with the config.ini.

highcutfloat, optional

The higher bound of the Filter. It is provided with the config.ini.

orderInt, optional

The order of the Filter. It is provided with the config.ini. Keep in mind that due to forward-backwards filtering the effektive order doubles.

resampledBool, optional

Wether the Signal is resampled or not. the resampled signal is a different attribute of this object. The default is False.

resampling(sampling_rate, target_rate=128, interpolate_pchip=False)

A function to resample a signal. It can be resampled with the C2-spline-interpolation method or the pchip method. This funktion is essentially deprecated but can still be used by modifying the PPG_EVA_tool.preprocessing funktion. The resampled signal and timestamps is stored in the self.resample_signal and self.new_timestamps atributes respectivley.

Parameters

sampling_rateInt

The original smaplingrate of the signal.

target_rate: Int

The target samplingrate of the signal. This parameter is provided by the config.ini as fs_A.

interpolate_pchip: bool

if True, the resampling is performed by the pchip-method. The default is False and hence C2-Splineinterpolation is used.

slice_list(chunk_size)

Sequencing of the signal into desired length. Before storing the signalchunks it is checked that all chunks are of the same length. The chunks are stored in the self.signal_chunks atribute as a 2D-array.

Parameters

chunk_sizeInt, optional

The desired length of sequence.

class eva_classes.SQI(signal_chunks)

Bases: object

The class to calculate all SQIs. To initialize it takes only a 2D-array of the signalchunks. Includes:´

shanon_entropy

calc_SNR

ZCR

skewness

kurt

ZCR()

Calculate the ZCR for every sequence. The ZCR is here defined as the variance to a lin. Regression of the time of occurance of every signchange. The lin. Regression is performed with the eva_toolkit.lin_reg() function and the variance is calculated by the eva_toolkit.variance() function.

Returns

self.signs: list

A list with values for the signs of the signalvalues.

1 = positive

-1 = negative

self.cross_pos: list

A list of the indices where a zero crossing occured.

self.n_cross: Int

The number of zero crossings.

self.slope: float

the slope of the lin. regression for every sequence.

self.intersect: float

the y-intersect of the lin. regression for every sequence.

self.variance:

the variance for every sequence.

calc_SNR(lowcut=0.8, highcut=2.0)

Calculates the Signal- to Noise-Ratio. Defined as the relative power of a signalband to the rest. To determine the signalpower, the FFT of the given chunk is integrated within the lowcut and highcut as lower and upper limits respectivley. The noisepower is the integrates value of the FFT outside these boundaries.

Parameters

lowcutfloat, optional

Lower bound of the frequency band. This parameter is provided by the config.ini as bpm.

highcutfloat, optional

higher bound of the frequency band. This parameter is provided by the config.ini as bpm.

Returns

self.SNR: list

The SNR-values for every sequence.

self.freq: list

the frequencies for every sequence.

self.magnitude: list

the magnitudes of the frequencies for every sequence.

kurt()

Calculates the kurtosis for every sequence.

Returns

self.kurt: list

The kurtosis for every sequence.

shanon_entropy(num_bins=16)

Calculates the normalised entropy for every sequence. The entropy for every chunk is stored in a array in the self.entropy_values atribute.

Parameters

num_binsInt, optional

The number of bins used to create the histogram and therefore the probability density function. The default is 16.

skewness()

Calculates the skewness for every sequence.

Returns

self.skewness: list

The skewness for every sequence.

class eva_classes.Training(skew_values, kurt_values, data)

Bases: object

The class for Training. It takes the skewness, and kurtosis as separate arrays and the anotation as a list or array as a binary classification.

Includes:

separate_values

building hists

building_hists(plot=False)

Builds a 2D probability densitifunction where the skewness and the kurtosis the two dimensions resemble. The resulting PDF is a logarithmic difference of the PDFs of the good data and the bad data.

Parameters

plotbool, optional

If True, all PDFs will be plotted. The default is False.

Returns

self.hist: aray

The PDF-values.

self.xedges: array

The xedges of the bins.

self.yedges: array

The yedges of the bins.

separate_values()

Separating the values in good and bad data according to the annotation.

Returns

self.positive_skew: array

All skewness values that are labeled as good.

self.positive_kurt: array

All kurtosis values that are labeled as good.

self.negative_skew: array

All skewness values that are labeled as bad.

self.negative_kurt: array

All kurtosis values that are labeled as bad.