maxfuse.graph.get_umap_embeddings

maxfuse.graph.get_umap_embeddings(arr, graph_rows, graph_cols, graph_vals, n_components=2, n_epochs=200, verbose=True)[source]

Compute UMAP embeddings.

Parameters:

  • arr (np.array of shape (n_samples, n_features)) – Data matrix

  • graph_rows (lists) – Each edge of the graph is graph_rows[i], graph_cols[i], and its weight is graph_vals[i]

  • graph_cols (lists) – Each edge of the graph is graph_rows[i], graph_cols[i], and its weight is graph_vals[i]

  • graph_vals (lists) – Each edge of the graph is graph_rows[i], graph_cols[i], and its weight is graph_vals[i]

  • n_components (int, default=2) – Number of dimensions of the embeddings

  • n_epochs (int, default=200) – Number of epochs to run

  • verbose (bool, default=True) – Whether to print the progress

Returns:

embeddings (array_like of shape (n_samples, n_components))