mlprimitives.candidates.dsp module

class mlprimitives.candidates.dsp.SpectralMask(method='std_dev', gain=1, window_length=128, beta=4)

Bases: object

Anomalies detection in satellite telemetry data using a spectral mask

fit(X)

Defines a spectral mask based on training data

Parameters

X – Training data

fit_freq_min_max(training_signal)

Defines a spectral mask based on training data using min and max values of each

frequency component

Parameters

training_signal – Training data

fit_freq_std_dev(training_signal)

Defines a spectral mask based on training data using the standard deviation values of

each frequency component

Parameters

training_signal – Training data

produce(X)

Detects anomalies in telemetry data based on its power spectral density

Parameters

X – Telemetry data

Returns

Data vector consisting of the anomalies detected in the telemetry data

Return type

anomalies

window_design(window_length, beta)

Kaiser window design

Parameters

• window_length – Length of the window in number of samples

• beta – Beta value for Kaiser window design

Returns

Window designed using the beta and length provided as inputs

Return type

window

mlprimitives.candidates.dsp.next_power_of_2(x)

Finds the next power of 2 value

Parameters

x – Input value

Returns

Next power of 2 value

Return type

power_of_2