OnlineFFTPeriodicityDetector#

class pyriodicity.OnlineFFTPeriodicityDetector(window_size, buffer_size=None, window_func='boxcar', detrend_func='linear')#

Online Fast Fourier Transform (FFT) based periodicity detector.

Detect periodicities in a signal stream data using Sliding Discrete Fourier Transform (DFT) algorithm [1].

Parameters:

window_sizeint: Size of the sliding window for the ACF computation.
buffer_sizeint, optional, default = 2 * window_size: Size of the samples buffer. Must be at least equal to window_size.
window_funcfloat, str, tuple, optional, default = ‘boxcar’: Window function to apply. See scipy.signal.get_window for accepted formats of the window parameter.
detrend_func{‘constant’, ‘linear’}, optional, default = ‘linear’: The kind of detrending to apply. If None, no detrending is applied.