cdlib.ensemble.pool_grid_filter

pool_grid_filter(graph: <Mock id='139903590627984'>, methods: Callable[[<Mock id='139903585207440'>, dict], object], configurations: list, quality_score: Callable[[<Mock id='139903585138320'>, object], object], aggregate: Callable[[list], object] = <built-in function max>) → tuple

Execute a pool of community discovery internal on the input graph. Returns the optimal partition for each algorithm given the specified quality function.

Parameters:
  • methods – list community discovery methods (from nclib.community)
  • graph – networkx/igraph object
  • configurations – list of lists (one for each method) of Parameter and BoolParameter objects
  • quality_score – a fitness function to evaluate the obtained partition (from nclib.evaluation)
  • aggregate – function to select the best fitness value. Possible values: min/max
Returns:

at each call the generator yields a tuple composed by: the actual method, its optimal configuration; the obtained communities; the fitness score.
Raises:

ValueError – if the number of methods is different from the number of configurations specified
Example: 
>>> import networkx as nx
>>> from cdlib import algorithms, ensemble
>>> g = nx.karate_club_graph()
>>> # Louvain
>>> resolution = ensemble.Parameter(name="resolution", start=0.1, end=1, step=0.1)
>>> randomize = ensemble.BoolParameter(name="randomize")
>>> louvain_conf = [resolution, randomize]
>>>
>>> # Angel
>>> threshold = ensemble.Parameter(name="threshold", start=0.1, end=1, step=0.1)
>>> angel_conf = [threshold]
>>>
>>> methods = [algorithms.louvain, algorithms.angel]
>>>
>>> for communities, scoring in ensemble.pool_grid_filter(g, methods, [louvain_conf, angel_conf], quality_score=evaluation.erdos_renyi_modularity, aggregate=max):
>>>     print(communities, scoring)