def analyze_cluster_run_diff(year, graphfile):
with multiprocessing.Pool(4, initializer=init_worker, initargs=(graphfile,)) as p:
clusterings = p.map(_cluster, range(100))
init_worker(graphfile)
scores_nmi = [
evaluation.normalized_mutual_information(
NodeClustering(c1, g, None), NodeClustering(c2, g, None)
).score
for c1, c2 in itertools.combinations(clusterings, 2)
]
scores_rand = [
evaluation.adjusted_rand_index(
NodeClustering(c1, g, None), NodeClustering(c2, g, None)
).score
for c1, c2 in itertools.combinations(clusterings, 2)
]
# scores_pairs = [
# evaluation.adjusted_mutual_information(
# NodeClustering(c1, g, None),
# NodeClustering(c2, g, None)
# ).score
# for c1, c2 in zip(clusterings[::2],clusterings[1::2])
# ]
return {
"NMI": scores_nmi,
"Rand": scores_rand,
}
def test_adjusted_rand(self):
g = nx.karate_club_graph()
lp_communities = label_propagation(g)
louvain_communities = louvain(g)
score = evaluation.adjusted_rand_index(louvain_communities, lp_communities)
self.assertLessEqual(score.score, 1)
self.assertGreaterEqual(score.score, 0)
def get_scores(idx1, idx2):
c1 = significant_clusterings[idx1]
c2 = significant_clusterings[idx2]
score_nmi = evaluation.normalized_mutual_information(
NodeClustering(c1, None, None), NodeClustering(c2, None, None)
).score
score_rand = evaluation.adjusted_rand_index(
NodeClustering(c1, None, None), NodeClustering(c2, None, None)
).score
return score_nmi, score_rand
def adjusted_rand_index(
self, clustering: Clustering) -> evaluation.MatchingResult:
"""
Rand index adjusted for chance.
The Rand Index computes a similarity measure between two clusterings
by considering all pairs of samples and counting pairs that are
assigned in the same or different clusters in the predicted and
true clusterings.
The raw RI score is then "adjusted for chance" into the ARI score
using the following scheme::
ARI = (RI - Expected_RI) / (max(RI) - Expected_RI)
The adjusted Rand index is thus ensured to have a value close to
0.0 for random labeling independently of the number of clusters and
samples and exactly 1.0 when the clusterings are identical (up to
a permutation).
ARI is a symmetric measure::
adjusted_rand_index(a, b) == adjusted_rand_index(b, a)
:param clustering: NodeClustering object
:return: ARI score
:Example:
>>> from cdlib.algorithms import louvain
>>> g = nx.karate_club_graph()
>>> communities = louvain(g)
>>> leiden_communities = algorithms.leiden(g)
>>> mod = communities.adjusted_rand_index(leiden_communities)
:Reference:
1. Hubert, L., & Arabie, P. (1985). **Comparing partitions**. Journal of classification, 2(1), 193-218.
"""
return evaluation.adjusted_rand_index(self, clustering)
print(testg)
fluid = nx.algorithms.community.asyn_fluidc(testg, 10, seed=5)
fluid2 = [list(x) for x in fluid]
fluid3 = cdlib.NodeClustering(fluid2, testg, "FluidWeight")
wcom = wf1.weight_fluid(testg, 10, seed=3)
wcoms = [list(x) for x in wcom]
wcoms2 = cdlib.NodeClustering(wcoms, testg, "FluidWeight")
w2com = wf2.asyn_fluidcWeight(testg, 10, seed=3)
w2coms = [list(x) for x in w2com]
w2coms2 = cdlib.NodeClustering(w2coms, testg, "FluidWeight")
print(evaluation.adjusted_rand_index(wcoms2, w2coms2))
# print("Alg1")
# print(evaluation.newman_girvan_modularity(testg, wcoms2).score)
# louvain = algorithms.louvain(testg, weight='weight', resolution=1.5)
reso = [14, 3.5, 1.5, 1, 0.7, 0.6, 0.4]
comNumb = [5, 7, 10, 13, 15, 17, 20]
# with open('modularityV3.csv', 'w', newline='') as file:
# writer = csv.writer(file)
# print("Louvain")
# writer.writerow(["05/06"])
# writer.writerow(["Louvain"])
# writer.writerow(["Communities", "Modularity"])
# for r, c in zip(reso, comNumb):