reservoir_computing.datasets

Attributes

downloader

Classes

`ClfLoader`	Class to download and load time series classification datasets.
`PredLoader`	Class to download and load time series forecasting datasets.
`SynthLoader`	Class to generate synthetic time series.

Functions

`mackey_glass`([sample_len, tau, delta_t, seed, n_samples])	Generate the Mackey Glass time-series.
`mso`([T, N, seed, freq])	Generates the Multiple Sinewave Oscillator (MSO) time-series
`lorenz`([sigma, rho, beta, y0, t_span, dt])	Generate the Lorenz attractor time-series.
`rossler`([a, b, c, y0, t_span, dt])	Generate the Rossler attractor time-series.

Module Contents

class reservoir_computing.datasets.ClfLoader

Class to download and load time series classification datasets.

datasets

available_datasets(details=False)

Print the available datasets.

Parameters:

detailsbool: If True, print a description of the datasets.

Returns:

None

get_data(alias)

Download and load the dataset.

Parameters:

aliasstr: The alias of the dataset to be downloaded.

Returns:

Xtrnp.ndarray: Training data
Ytrnp.ndarray: Training labels
Xtenp.ndarray: Test data
Ytenp.ndarray: Test labels

class reservoir_computing.datasets.PredLoader

Class to download and load time series forecasting datasets.

datasets

available_datasets(details=False)

Print the available datasets.

Parameters:

detailsbool: If True, print a description of the datasets.

Returns:

None

get_data(alias) → numpy.ndarray

Download and load the dataset.

Parameters:

aliasstr: The alias of the dataset to be downloaded.

Returns:

Xnp.ndarray: Time series data

reservoir_computing.datasets.mackey_glass(sample_len=1000, tau=17, delta_t=1, seed=None, n_samples=1)

Generate the Mackey Glass time-series.

Parameters:

sample_lenint (default 1000): Length of the time-series in timesteps.
tauint (default 17): Delay of the MG system. Commonly used values are tau=17 (mild chaos) and tau=30 (moderate chaos).
delta_tint (default 1): Time step of the simulation.
seedint or None (default None): Seed of the random generator. Can be used to generate the same timeseries each time.
n_samplesint (default 1): Number of samples to generate.

Returns:

np.ndarray | list: Generated Mackey-Glass time-series. If n_samples is 1, a single array is returned. Otherwise, a list.

reservoir_computing.datasets.mso(T=1000, N=10, seed=None, freq=0.5)

Generates the Multiple Sinewave Oscillator (MSO) time-series by combining inusoids with incommensurable periods. The sinusoids to be combined are selected randomly.

Parameters:

Tint (default 1000): Number of time steps.
Nint (default 10): Maximum number of sinusoids to combine.
seedint or None (default None): Seed for the random generator.
freqfloat (default 0.5): Frequency of the sinusoids.

Returns:

np.ndarray: MSO time-series.

reservoir_computing.datasets.lorenz(sigma=10, rho=28, beta=8 / 3, y0=[0, -0.01, 9.0], t_span=[0, 100], dt=0.001)

Generate the Lorenz attractor time-series.

Parameters:

sigmafloat (default 10): 1st parameter of the Lorenz system.
rhofloat (default 28): 2nd parameter of the Lorenz system.
betafloat (default 8/3): 3rd parameter of the Lorenz system.
y0list (default [0, -0.01, 9.0]): Initial conditions of the Lorenz system.
t_spanlist (default [0, 100]): Time span of the simulation.
dtfloat (default 1e-3): Time step of the simulation.

Returns:

np.ndarray: Lorenz time-series.

reservoir_computing.datasets.rossler(a=0.2, b=0.2, c=5.7, y0=[0.5, 0.5, 0.5], t_span=[0, 200], dt=0.001)

Generate the Rossler attractor time-series.

Parameters:

afloat (default 0.2): 1st parameter of the Rossler system.
bfloat (default 0.2): 2nd parameter of the Rossler system.
cfloat (default 5.7): 3rd parameter of the Rossler system.
y0list (default [0, 0.1, 0]): Initial conditions of the Rossler system.
t_spanlist (default [0, 100]): Time span of the simulation.
dtfloat (default 1e-3): Time step of the simulation.

Returns:

np.ndarray: Rossler time-series.

class reservoir_computing.datasets.SynthLoader

Class to generate synthetic time series.

datasets

available_datasets(details=False): Print the available synthetic datasets.

Returns:

None

get_data(alias, **kwargs)

Generate the synthetic time series.

Parameters:

aliasstr: The alias of the synthetic dataset to be generated.
kwargsdict: Additional parameters for the synthetic dataset.

Returns:

np.ndarray: Synthetic time series.

reservoir_computing.datasets.downloader