-
Notifications
You must be signed in to change notification settings - Fork 0
/
edgeCode2Community.m
143 lines (126 loc) · 7.83 KB
/
edgeCode2Community.m
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
function community = edgeCode2Community(edgeCode,edgeAdjMatrix,edgeMatrix,AdjMatrix,similarity,M)
community = {};
nodeNum =size(AdjMatrix,1);
f_node = zeros(1,nodeNum);
edgeNum = size(edgeMatrix,1);
[ff,index] = sort(edgeCode);
label_index = [ff;index];%第一行为边标签值，第二行为该标签所对应边在边矩阵里的索引
labels = unique(edgeCode);
labelsNum = size(labels,2);
labels_nodeSet = zeros(labelsNum,nodeNum);%存放边标签所对应的点集合
labels_advSim = zeros(labelsNum,2);%存放边标签所对应的边的平均相似度和边数
% % 解码
if(labelsNum==edgeNum)%一个边一个划分，解码成一个点一个划分
f_node(1,:) = 1:nodeNum;
else
for label_i = 1:labelsNum
edgeIndex = label_index(2,find(label_index(1,:) == labels(label_i)));
%计算该集合边之间的平均相似度
sumSim = 0;
% nchoosek(edgeIndex(1,:),2);
for sim_i = 1:size(edgeIndex,2)
for sim_j = (sim_i+1):size(edgeIndex,2)
sumSim = sumSim+similarity(sim_i,sim_j);
end
end
labels_advSim(label_i,:) = [sumSim/size(edgeIndex,2),size(edgeIndex,2)];
for label_j = 1:size(label_index,2)
if(labels(label_i)==label_index(1,label_j))
edge = edgeMatrix(label_index(2,label_j),:);%由边索引找到边所包含的顶点
% f_node(edge(1)) = labels(label_i);%将边的标签值转化为对应点的标签值
% f_node(edge(2)) = labels(label_i);
%labels_nodeSet(label_i,1) = labels(label_i);
labels_nodeSet(label_i,edge) = edge;%第i行表示第i个边标签包含的节点
end
end
end
%%化成community胞元数组
% for ci = 1:size(labels_nodeSet,1)
% community{ci} = labels_nodeSet(ci,find(labels_nodeSet(ci,:)~=0));
% end
for labels_nodeSet_i = 1:size(labels_nodeSet,1)
for labels_nodeSet_j = (labels_nodeSet_i+1):size(labels_nodeSet,1)
intersect_node = intersect(labels_nodeSet(labels_nodeSet_i,:),labels_nodeSet(labels_nodeSet_j,:));
union_node = union(labels_nodeSet(labels_nodeSet_i,:),labels_nodeSet(labels_nodeSet_j,:));
intersect_node(intersect_node==0) = [];
union_node(union_node==0) = [];
for k = 1:length(intersect_node)
set1 = labels_nodeSet(labels_nodeSet_i,:);
set2 = labels_nodeSet(labels_nodeSet_i,:);
f_in1 = sum(sum(AdjMatrix(find(set1(1,:)~=0),find(set1(1,:)~=0))))/2;
f_out1 = sum(sum(AdjMatrix(find(set1(1,:)~=0),find(set1(1,:)==0))));
set2(1,find(set1(1,:)==intersect_node(1,k)))=0;%社团A除去候选重叠点后的点集合
f_in2 = sum(sum(AdjMatrix(find(set2(1,:)~=0),find(set2(1,:)~=0))))/2;
f_out2 = sum(sum(AdjMatrix(find(set2(1,:)~=0),find(set2(1,:)==0))));
Q1 = f_in1/(f_in1+f_out1);
Q2 = f_in2/(f_in2+f_out2);
if Q1 <= Q2%模块度变小，删除该候选重叠点
labels_nodeSet(labels_nodeSet_i,find(labels_nodeSet(labels_nodeSet_i,:)==intersect_node(1,k)))=0;
end
set11 = labels_nodeSet(labels_nodeSet_j,:);
set22 = labels_nodeSet(labels_nodeSet_j,:);
f_in11 = sum(sum(AdjMatrix(find(set11(1,:)~=0),find(set11(1,:)~=0))))/2;
f_out11 = sum(sum(AdjMatrix(find(set11(1,:)~=0),find(set11(1,:)==0))));
set22(1,find(set11(1,:)==intersect_node(1,k)))=0;%社团A除去候选重叠点后的点集合
f_in22 = sum(sum(AdjMatrix(find(set22(1,:)~=0),find(set22(1,:)~=0))))/2;
f_out22 = sum(sum(AdjMatrix(find(set22(1,:)~=0),find(set22(1,:)==0))));
Q11 = f_in11/(f_in11+f_out11);
Q22 = f_in22/(f_in22+f_out22);
if Q11 <= Q22%模块度变小，删除该候选重叠点
labels_nodeSet(labels_nodeSet_j,find(labels_nodeSet(labels_nodeSet_j,:)==intersect_node(1,k)))=0;
end
end
if abs(labels_advSim(labels_nodeSet_i,1)-labels_advSim(labels_nodeSet_j,1))>=0.15%比较两个标签的平均相似度
%将相似度较小的标签点集合中将公共部分剔除
if labels_advSim(labels_nodeSet_i,1)>labels_advSim(labels_nodeSet_j,1)
setdiffSet = setdiff(labels_nodeSet(labels_nodeSet_j,:),intersect_node);
setdiffSet(setdiffSet==0)=[];
labels_nodeSet(labels_nodeSet_j,:) = 0;
labels_nodeSet(labels_nodeSet_j,setdiffSet) = setdiffSet;%只保留差集
elseif labels_advSim(labels_nodeSet_j,1)>labels_advSim(labels_nodeSet_i,1)
setdiffSet = setdiff(labels_nodeSet(labels_nodeSet_i,:),intersect_node);
setdiffSet(setdiffSet==0)=[];
labels_nodeSet(labels_nodeSet_i,:) = 0;
labels_nodeSet(labels_nodeSet_i,setdiffSet) = setdiffSet;%只保留差集
end
% elseif labels_advSim(labels_nodeSet_i,2)>labels_advSim(labels_nodeSet_j,2)%比较两个标签的节点数目
% setdiffSet = setdiff(labels_nodeSet(labels_nodeSet_j,:),intersect_node);
% setdiffSet(setdiffSet==0)=[];
% labels_nodeSet(labels_nodeSet_j,:) = 0;
% labels_nodeSet(labels_nodeSet_j,setdiffSet) = setdiffSet;%只保留差集
% elseif labels_advSim(labels_nodeSet_i,2)<=labels_advSim(labels_nodeSet_j,2)%比较两个标签的节点数目
% setdiffSet = setdiff(labels_nodeSet(labels_nodeSet_i,:),intersect_node);
% setdiffSet(setdiffSet==0)=[];
% labels_nodeSet(labels_nodeSet_i,:) = 0;
% labels_nodeSet(labels_nodeSet_i,setdiffSet) = setdiffSet;%只保留差集
% else
% if rand(1)>=.5
% setdiffSet = setdiff(labels_nodeSet(labels_nodeSet_j,:),intersect_node);
% setdiffSet(setdiffSet==0)=[];
% labels_nodeSet(labels_nodeSet_j,:) = 0;
% labels_nodeSet(labels_nodeSet_j,setdiffSet) = setdiffSet;%只保留差集
% else
% setdiffSet = setdiff(labels_nodeSet(labels_nodeSet_i,:),intersect_node);
% setdiffSet(setdiffSet==0)=[];
% labels_nodeSet(labels_nodeSet_i,:) = 0;
% labels_nodeSet(labels_nodeSet_i,setdiffSet) = setdiffSet;%只保留差集
% end
end
end
end
end
labels_nodeSet(find(sum(abs(labels_nodeSet),2)==0),:)=[];%除去全零的行
for labels_nodeSet_j = 1:size(labels_nodeSet,1)
% if labels_nodeSet(labels_nodeSet_j,nodeLabel_i)~=0
community{labels_nodeSet_j} = labels_nodeSet(labels_nodeSet_j,find(labels_nodeSet(labels_nodeSet_j,:)~=0));
% end
end
% for nodeLabel_i = 1:nodeNum
% for labels_nodeSet_j = 1:size(labels_nodeSet,1)
% if labels_nodeSet(labels_nodeSet_j,nodeLabel_i)~=0
% f_node(1,nodeLabel_i) = labels_nodeSet_j;
% end
% end
% end
% node_code = f_node;
end