def test_Rossa():
# Rossa =========================
G=nx.karate_club_graph()
print("Running Rossa's algorithm ...")
rs = cpa.Rossa()
rs.detect(G)
pair_id = rs.get_pair_id()
coreness = rs.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, rs)
def test_KM_config():
# KM--config =========================
G=nx.karate_club_graph()
print("Running KM_config ...")
km = cpa.KM_config()
km.detect(G)
pair_id = km.get_pair_id()
coreness = km.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, km)
def test_BE():
# BE =========================
G=nx.karate_club_graph()
print("Running BE algorithm ...")
be = cpa.BE()
be.detect(G)
pair_id = be.get_pair_id()
coreness = be.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, be)
def test_LapSgnCore():
# LapSgnCore =========================
G=nx.karate_club_graph()
print("Running LapSgnCore algorithm ...")
lsc = cpa.LapSgnCore()
lsc.detect(G)
pair_id = lsc.get_pair_id()
coreness = lsc.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, lsc)
def test_SBM():
# SBM =========================
G=nx.karate_club_graph()
print("Running SBM algorithm ...")
sbm = cpa.SBM()
sbm.detect(G)
pair_id = sbm.get_pair_id()
coreness = sbm.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, sbm)
def test_MINRES():
# MINRES =========================
G=nx.karate_club_graph()
print("Running MINRES algorithm ...")
mrs = cpa.MINRES()
mrs.detect(G)
pair_id = mrs.get_pair_id()
coreness = mrs.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, mrs)
def test_Surprise():
# Surprise =========================
G=nx.karate_club_graph()
print("Running Surprise ...")
spr = cpa.Surprise()
spr.detect(G)
pair_id = spr.get_pair_id()
coreness = spr.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, spr)
def test_Divisive():
# KM--config =========================
G=nx.karate_club_graph()
print("Running Divisive ...")
dv = cpa.Divisive()
dv.detect(G)
pair_id = dv.get_pair_id()
coreness = dv.get_coreness()
sig_pair_id, sig_coreness, significance, p_values = cpa.qstest(pair_id, coreness, G, dv)
#only needed for continues models like Rossa
cut = 0.15
algorithm = cp.MINRES()
algorithm.detect(DPI)
#dictionary of nodes, telling to which core-periphery pair the node belongs (if only single core
#-periphery structure then then each key has the value 0)
c = algorithm.get_pair_id()
#Dictionary with node as keys telling if the node is a core node (1) or periphery node(0)
x = algorithm.get_coreness()
# this function calculates the significance that the network is a core perihpery network by comparing it to
# 1000 random networks (configuration network)
sig_c, sig_x, significant, p_values = cp.qstest(c, x, DPI, algorithm, num_of_rand_net=500)
#sig_c same as c but only for significant assiciations (else None value)
#sig_x same as x but only for significant assiciations (else None value)
#boolean True/False if network is core perihpery structures
#p_values for the individual core-periphery structures
print(sig_x)
print('Significant Core-Periphery Structure?')
print(significant)
print('Pvalues: ')
print(p_values)
if model != 'Rossa':
print('Core Nodes: %d' %len([node for node in sig_x if sig_x[node] == 1]))
