Multi-view latent Gaussian dataset

A synthetic multi-view dataset where each view is generated from a shared latent Gaussian distribution.

polyview.dataset.multiview_latent_gaussian.generate_multiview_gaussian(n_samples=300, n_features=10, latent_dim=3, centers=3, n_views=2, noise_std=0.1, random_state=None)

Generate a synthetic multi-view dataset using a shared latent Gaussian model.

Parameters:

  • n_samples (int, default=300) – Number of samples.

  • n_features (int, default=10) – Number of features per view.

  • latent_dim (int, default=3) – Dimension of shared latent space.

  • centers (int, default=3) – Number of clusters in latent space.

  • n_views (int, default=2) – Number of views.

  • noise_std (float, default=0.1) – Standard deviation of Gaussian noise.

  • random_state (int or None, default=None) – Seed.

Returns:

A MultiViewDataset object containing the generated views and labels.

Return type:

MultiViewDataset