make_sinusoidal_process#

skfda.datasets.make_sinusoidal_process(n_samples=15, n_features=100, *, start=0, stop=1, period=1, phase_mean=0, phase_std=0.6, amplitude_mean=1, amplitude_std=0.05, error_std=0.2, random_state=None)[source]#

Generate sinusoidal proccess.

Each sample \(x_i(t)\) is generated as:

\[x_i(t) = \alpha_i \sin(\omega t + \phi_i) + \epsilon_i(t)\]

where \(\omega=\frac{2 \pi}{\text{period}}\). Amplitudes \(\alpha_i\) and phases \(\phi_i\) are normally distributed. \(\epsilon_i(t)\) is a gaussian white noise process.

Parameters:
  • n_samples (int) – Total number of samples.

  • n_features (int) – Points per sample.

  • start (float) – Starting point of the samples.

  • stop (float) – Ending point of the samples.

  • period (float) – Period of the sine function.

  • phase_mean (float) – Mean of the phase.

  • phase_std (float) – Standard deviation of the phase.

  • amplitude_mean (float) – Mean of the amplitude.

  • amplitude_std (float) – Standard deviation of the amplitude.

  • error_std (float) – Standard deviation of the gaussian Noise.

  • random_state (int | RandomState | Generator | None) – Random state.

Returns:

FDataGrid object comprising all the samples.

Return type:

FDataGrid

Examples using skfda.datasets.make_sinusoidal_process#

Extrapolation

Extrapolation

Interpolation

Interpolation

Radius neighbors classification

Radius neighbors classification

Shift Registration

Shift Registration