forked from ketkik22/Community-detection-on-social-network
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CommunityDetectionUsingKruskalsAlgorithm.py
195 lines (157 loc) · 5.76 KB
/
CommunityDetectionUsingKruskalsAlgorithm.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
import operator
from igraph import *
import queue as queue
from unionfind import unionfind
from random import randint
"""
Functions
"""
def calculateWeightOfEdgeUsingNeighborhoodOverlap(graph, edge):
"This fucntion will calculate the neighborhood overlap of an edge and return it"
NP = 0
neighborsA = graph.neighbors(edge.source)
neighborsB = graph.neighbors(edge.target)
commonNeighbors = 0
if len(neighborsA) < len(neighborsB):
for vertex in neighborsA:
if vertex in neighborsB:
commonNeighbors += 1
else:
for vertex in neighborsB:
if vertex in neighborsA:
commonNeighbors += 1
degreeA = graph.degree(edge.source,loops=False)
degreeB = graph.degree(edge.target,loops=False)
if((degreeA + degreeB - 2) == 0):
NP = 0
else:
NP = commonNeighbors / (degreeA + degreeB - 2)
return NP
def calculateWeightOfEdgeUsingDegree(graph, edge):
"This method calculates the weight of the edge using degree of its endpoints"
weight = len(graph.neighbors(edge.source)) + len(graph.neighbors(edge.target))
return weight
"""
Visual Style
"""
visual_style = {}
visual_style["edge_curved"] = False
visual_style["vertex_color"] = "#13a53f"
visual_style["vertex_size"] = 16
visual_style["vertex_label_size"] = 12
visual_style["vertex_label_color"] = "#ffffff"
visual_style["bbox"] = (300, 230)
"""
Input graph
"""
x = 18
#inputGraph = Graph.Read_GraphML('karate.GraphML')
#inputGraph = Graph.Read_Edgelist('0.edges',directed=False)
#inputGraph = read('football.gml')
#inputGraph = read('dolphins.gml')
inputGraph = Graph.Barabasi(n=x, m=3, zero_appeal=3)
print(summary(inputGraph))
#inputGraph.vs["label"] = range(1000)
for v in inputGraph.vs():
v['label'] = v.index
for edge in inputGraph.es():
#weight = calculateWeightOfEdgeUsingNeighborhoodOverlap(inputGraph, edge)
weight = calculateWeightOfEdgeUsingDegree(inputGraph, edge)
edge["weight"] = weight
#edge['label'] = NP
plot(inputGraph, "graph01.png", **visual_style)
#print summary(g)
"""
Existing community detection algorithms: Louvain & Girvan-Newman
"""
louvainCommunity = inputGraph.community_multilevel()
plot(louvainCommunity, "louvain community.png", mark_groups=True)
print("\n************************* LOUVAIN ALGORITHM *************************")
print(summary(louvainCommunity))
print("Modularity of Louvain algorithm = %f\t\t"%louvainCommunity.q,"\n\n")
girvanNewmanCommunity = inputGraph.community_edge_betweenness().as_clustering()
plot(girvanNewmanCommunity, "girvan-newman community.png", mark_groups=True)
print("************************* GIRVAN-NEWMAN ALGORITHM ********************")
print(summary(girvanNewmanCommunity))
print("Modularity of Girvan-Newman algorithm = %f"%girvanNewmanCommunity.q,"\n\n\n")
mstTree = Graph.spanning_tree(inputGraph, weights=inputGraph.es['weight'], return_tree=True)
plot(mstTree, "spanningTreeByPrim.png", **visual_style)
mst = inputGraph.copy()
mst.delete_edges(mst.es())
"""
Running kruskal's algorithm
"""
priotrityQWeights = queue.PriorityQueue()
unionDS = unionfind(inputGraph.vcount())
alreadyUsedEdges = []
for edge in inputGraph.es():
#print(type(edge))
weight = edge["weight"]
priotrityQWeights.put(weight)
#print(edge.tuple, "-------", edge['weight'])
while not priotrityQWeights.empty():
weight = priotrityQWeights.get()
#print(priotrityQWeights.qsize())
for e in inputGraph.es():
if (e['weight'] == weight and e not in alreadyUsedEdges):
edge = e
alreadyUsedEdges.append(edge)
break
#print(edge.tuple, "-------", edge['weight'])
if(unionDS.find(edge.source) != unionDS.find(edge.target)):
unionDS.unite(edge.source, edge.target)
mst.add_edges([(edge.source, edge.target)])
plot(mst, "MST.png", **visual_style)
"""
calculating edge betweenness
"""
ebList = mst.edge_betweenness()
alreadyUsedValuesOfK = []
mst1 = mst.copy()
"""
Approximation of value of k
"""
count = 0
index = 0
maxModularity = -1
while count < 20:
while len(alreadyUsedValuesOfK) != mst.ecount():
k = randint(1, mst.ecount())
if k not in alreadyUsedValuesOfK:
alreadyUsedValuesOfK.append(k)
break
print ("************************************************ k = ", k)
mst = mst1.copy()
ebList = mst.edge_betweenness()
"""
Removing k-1 edges
"""
i = 0
while i < k-1:
tupleId=0
maxEdgeBetweenness = max(ebList)
for idx, eb in enumerate(ebList):
if(maxEdgeBetweenness == eb):
tupleId = idx;
break;
#if(mst.are_connected(mst.es[tupleId].tuple[0], mst.es[tupleId].tuple[1])):
if(tupleId < mst.ecount()):
mst.delete_edges([(mst.es[tupleId].tuple[0], mst.es[tupleId].tuple[1])])
ebList.remove(maxEdgeBetweenness)
i+= 1
comm = mst.clusters(mode="STRONG")
modularity = comm.q
if (maxModularity < modularity):
maxModularity = modularity
index = k
maxModularityCommunities = comm
print(summary(comm))
print("Modularity = %f\t\t" % (comm.q))
# print("%d Modularity with igraph method = %f\t\t" % (idx, componentList.q),summary(vertexCluster))
plot(comm, "outputGraph%d.png" %k, mark_groups=True)
count += 1
print("\n\nMax modularity index = %d"%index)
print("Number of communities = ", len(set(maxModularityCommunities.membership)))
print("Modularity = %f\t\t" % (maxModularity))
plot(maxModularityCommunities, "outputGraph.png", mark_groups=True)
exit()