pyagc.metrics.conductance

conductance(edge_index: Tensor, clusters: Tensor, vectorized: bool = True) → float

Computes the average conductance of clusters in a graph. This implementation is adapted from: google-research/dmon.

Parameters:

• edge_index (torch.Tensor) – (2, num_edges), undirected edge list.

• clusters (torch.Tensor) – (num_nodes,), cluster assignments.

• vectorized (bool, optional) – Whether to use vectorized computation. (default: True)

Returns:

float – Average conductance of all clusters.