Module: cluster

In this module are all the functions to identify clusters in a given contact map or evaluate data if clustering makes sense (Hopkins statistic). A cluster is a set of contacts which are directly or indirectly (chain of direct contacts) connected.

Members

BioHelpers_FABER.cluster.findCluster(x: ndarray) → list

Identifying all clusters.

Parameters:

x (np.array) – contact map

Returns:

list of clusters [cluster1, cluster2, …] where clusterX of the form [[x1,y1],[x2,y2],…]

Return type:

list

BioHelpers_FABER.cluster.hopkins(X: ndarray, subsample: float = 0.1, seed: int = 247) → float

Hopkins Function

Parameters:

• X (np.ndarray) – Datapoints with dimension (n,d)

• subsample (float, optional) – Fraction of subsample, defaults to 0.1

• seed (int, optional) – Seed for reproducibility, defaults to 247

Returns:

Hopkins Value

Return type:

float