uadapy.data package

uadapy.data.data module

uadapy.data.data.generate_synthetic_gmm(n_classes=3, n_dims=4, random_state=0)

Generates synthetic Gaussian Mixture Model distributions. Creates multiple classes, each represented as a GMM with random number of components (1 to 10). Per component, random means and covariances are generated to form the GMM.

Parameters:

  • n_classes (int, optional) – Number of classes to generate. Default value is 3.

  • n_dims (int, optional) – Dimensionality of the original data space. Default value is 4.

  • random_state (int, optional) – Random seed for reproducibility. Default value is 0.

Returns:

List of Distribution objects, each wrapping a MultivariateGMM model.

Return type:

list

uadapy.data.data.generate_synthetic_timeseries(timesteps=200, trend=0.1)

Generates synthetic time series data by modeling a combination of trend, periodic patterns, and noise using a multivariate normal distribution with an exponential quadratic kernel for covariance.

Parameters:

timesteps (int) – The time steps of the time series. Default value is 200.

Returns:

timeseries – An instance of the TimeSeries class, which represents a univariate time series.

Return type:

Timeseries object

uadapy.data.data.load_iris()

Uses the iris dataset and fits a normal distribution :return:

uadapy.data.data.load_iris_gmm(n_components=2, random_state=0)

Uses the iris dataset and fits a Gaussian Mixture Model for each class.

Parameters:

  • n_components (int, optional) – Number of mixture components for each GMM. Default value is 2.

  • random_state (int, optional) – Random seed for reproducibility. Default value is 0.

Returns:

List of Distribution objects, each wrapping a MultivariateGMM model fitted to one class of the iris dataset.

Return type:

list

uadapy.data.data.load_iris_normal()

Uses the iris dataset and fits a normal distribution :return: