andrey.data.RealDataset

class andrey.data.RealDataset(data, graph, feature_names, description)

A dataset from load_dataset(): its observations and the graph they are scored against.

data

The observations, one row per sample and one column per variable. A DataFrame’s column names become the node labels of a graph learned from it. With as_frame=False, load_dataset() gives a float64 array in the same column order.

Type:

pandas.DataFrame or numpy.ndarray of shape (n_samples, n_variables)

graph

The reference DAG (kind="dag"), its nodes labeled with feature_names.

Type:

GraphStructure

feature_names

The variable names, in column order.

Type:

tuple of str

description

What the data are, whether they are measured or simulated, where they come from, and how to cite them.

Type:

str