ash_model.measures.attribute_analysis.attribute_consistency¶

ash_model.measures.attribute_analysis.attribute_consistency(h, node=None)[source]¶

The consistency measures the extent to which a nodes’ attribute value remains constant/change across time. The higher the value, the more consistent in time are the node’s values for that attribute.

Parameters:

h (ASH) – ASH instance
node (int | None) – Specify the node for which we want to calculate the consistency

Returns:

A dict containing, for each node, for each attribute, the consistency value

Return type:

dict

Examples

Consistency for a specific node over time (avoid node=0 which is falsy in current implementation):

>>> import numpy as np, networkx as nx
>>> from ash_model.utils.networkx import from_networkx_maximal_cliques_list
>>> Gs = [nx.barabasi_albert_graph(100, 3, seed=i) for i in range(10)]
>>> rng = np.random.default_rng(42)
>>> for G in Gs:
...     for n in G.nodes():
...         G.nodes[n]['color'] = 'red' if rng.integers(0, 2) == 0 else 'blue'
>>> h = from_networkx_maximal_cliques_list(Gs)
>>> attribute_consistency(h, node=1)
{'color': np.float64(0.1187091007693073)}

Note: passing node=0 currently behaves as if no node was specified because 0 evaluates to False.