recap descriptive analysis of network graph connectivity · 2014. 10. 9. · recap introduction...
TRANSCRIPT
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Descriptive Analysis of Network GraphCharacteristics
Network Analysis: Lecture 3
Sacha Epskamp
University of AmsterdamDepartment of Psychological Methods
16-09-2014
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A common method for constructing networks is by takingsome measure of similarity or association, sim(u, v)between each pair of nodes u, v ∈ V and either using thatsimilarity as edge weights for a weighted graph:
wuv = sim(u, v)
or only connecting two nodes in an unweighted graph iftheir similarity is not zero:
auv =
{1 if sim(u, v) 6= 00 if sim(u, v) = 0
In the lecture we used the correlation coefficient asassociation measure.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
1
2
34
5
6
7
89
10
11
12
1314
15
16
17
1819
20
21
22
2324
25
26
27
2829
30
31
32
3334
35
36
37
3839
40
41
42
4344
45
46
47
4849
50
51
52
5354
55
56
57
5859
60
61
62
6364
65
66
67
6869
70
71
72
7374
75
76
77
7879
80
81
82
8384
85
86
87
8889
90
91
92
9394
95
96
97
9899
100
101
102
103104
105
106
107
108109
110
111
112
113114
115
116
117
118119
120
121
122
123124
125
126
127
128129
130
131
132
133134
135
136
137
138139
140
141
142
143144
145
146
147
148149
150
151
152
153154
155
156
157
158159
160
161
162
163164
165
166
167
168169
170
171
172
173174
175
176
177
178179
180
181
182
183184
185
186
187
188189
190
191
192
193194
195
196
197
198199
200
201
202
203204
205
206
207
208209
210
211
212
213214
215
216
217
218219
220
221
222
223224
225
226
227
228229
230
231
232
233234
235
236
237
238239
240
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
Cutoff: 0.4Minimum: 0.25 Maximum: 0.77
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
● ●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
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
8283
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
176177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
Cutoff: 0.4Minimum: 0.25 Maximum: 0.77
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
To interpret qgraph networks, three values need to beknown:
Minimum Edges with absolute weights under thisvalue are omitted
Cut If specified, splits scaling of width and colorMaximum If set, edge width and color scale such that
an edge with this value would be the widestand most colorful
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
If the one-factor model is true, then the covariancebetween two variables equals the product of factorloadings!
I Correlations are non-zero if and only if factorloadings are non-zero
I The correlation-network of data generated by asingle factors portrays a fully connected cluster ofnodes:
y1
y2
y3
y4
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
If variable i has a stronger covariance than variable j onsome third variable and the one-factor model is true, i hasa stronger covariance than j to all other variables.
y1
y2
y3
y4
In line with one−factor model
y1
y2
y3
y4
Violates one−factor model
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Descriptive Analysis of Network GraphCharacteristics
Network Analysis: Lecture 3
Sacha Epskamp
University of AmsterdamDepartment of Psychological Methods
16-09-2014
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Prevent the outbreak:http://vax.herokuapp.com/
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
I How fast will the disease spread?I Connectivity
I What nodes should be vaccinated?I Centrality
I Which parts of the network should be quarantined?I Clustering
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
We will first analyze only the unweighted, undirectedsimple graph G:
G = (V ,E)
With |V | nodes and |E | edges, encoded using |V | × |V |adjacency matrix A:
auv = avu =
{1 if {u, v} ∈ E0 Otherwise
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The shortest path length between nodes v and u,dist(v ,u), is defined in an unweighted graph as theminimum number of steps you need to take from node vto reach node u:
dist(v ,u) = min (avx + . . .+ ayu)
I Can be computed using Dijkstra’s algorithm (Dijkstra,1959) with weights fixed to 1.
I Commonly referred to as the shortest path lengthor geodesic distance
The mean shortest path length is called the averageshortest path length (APL) and is an important measurefor how well connected a graph is:
APL(G) =
∑v ,u dist(v ,u)
|V | (|V | − 1)/2
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
shortest.paths(G)
## [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]## [1,] 0 1 1 1 1 1 2 2## [2,] 1 0 2 2 2 2 3 3## [3,] 1 2 0 2 2 2 3 3## [4,] 1 2 2 0 2 2 3 3## [5,] 1 2 2 2 0 2 1 1## [6,] 1 2 2 2 2 0 1 1## [7,] 2 3 3 3 1 1 0 2## [8,] 2 3 3 3 1 1 2 0## [9,] 2 3 3 3 1 1 2 2## [10,] 2 3 3 3 1 1 2 2## [11,] 2 3 3 3 1 1 2 2## [,9] [,10] [,11]## [1,] 2 2 2## [2,] 3 3 3## [3,] 3 3 3## [4,] 3 3 3## [5,] 1 1 1## [6,] 1 1 1## [7,] 2 2 2## [8,] 2 2 2## [9,] 0 2 2## [10,] 2 0 2## [11,] 2 2 0
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
average.path.length(G)
## [1] 2
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
●
●
● ●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●● ●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
● ●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●●
● ● ●
●●
●●●●
●
●●
●
●
●
●●
●●
●
● ●
●
●●
●
●
●
●
●●●
●●
●●
●
●●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●●
●●
●
●●
●
●●
●●
●
●
●
●●
●●
●
●●●
●
●
●
●
●●● ●
●●
●●●
●
●
●
●
●● ● ●
●
●
●
●
●
●●
●
●
●●
●●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
● ●
●
●
●
●●
●
●
●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●●
● ●●
● ●
●●
● ●
●●
●
●
●●
●
●●
●
● ●●
●
●●
●
●
●●
●
●●
●●
●
●●
● ●●
● ●
●
●
●●●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
3.5
3
2.5
2
1.5
1
MDE x
DYS
MDE x
AGPH
MDE x
SOP
MDE x
SIP
MDE x
PD
MDE x
APD
DYS X A
GPH
DYS X S
OP
DYS X S
IP
DYS X P
D
DYS X A
PD
AGPH
X S
OP
AGPH
X S
IP
AGPH
X P
D
AGPH
X A
PD
SOP X
SIP
SOP X
PD
SOP X
APD
SIP X
PD
SIP X
APD
PD X
APD
Correlation Average Shortest Path Length
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The diameter of a graph is its longest shortest pathlength:
diameter(G) = maxu,v
[dist(u, v)]
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
diameter(G)
## [1] 3
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The density of a graph is the proportion of the presentnumber of edges to the total possible amount of edges:
den(G) =|E |
|V | (|V | − 1)/2
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
graph.density(G)
## [1] 0.2727
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Are two connected nodes also connected to each other?Or more general, does a graph exhibit cliques?
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The local clustering coefficient, cl(v), gives for node vthe proportion of neighbors of v that are also connectedto each other.This corresponds for dividing the amount of “triangles” ofwhich node v is part, τ∆(v) to the amount of possibletriangles of which v could be part: τ3(v):
cl(v) =τ∆(v)τ3(v)
τ3(v) also corresponds to the number of triplets in whichnode v is the middle node.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Triangle Triplet
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
1
1
1
0.43
1
1
1
1
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
I The local clustering coefficient can also be seen as ameasure for redundancy
I A node with clustering of 1 has connected neighbors.Deleting this node will not hugely change thestructure of a network
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Traditionally, a global clustering coefficient for thewhole graph can be obtained by averaging all the localclustering coefficients:
cl(G) =1|V |
|V |∑i=1
cl(i)
This is an average of averages, and should be properlyweighted to obtain a more informative coefficient:
clT (G) =
∑|V |i=1 τ∆(i)cl(i)τ3(i)
=3τ∆(G)
τ3(G)
This is the more modern clustering coefficient, alsotermed transitivity
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
1
1
1
0.43
1
1
1
1
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A <- matrix(0,9,9)A[1:5,1:5] <- 1A[5:9,5:9] <- 1
library("igraph")
G4 <- graph.adjacency(A, mode = "undirected", diag = FALSE)
transitivity(G4,"local")
## [1] 1.0000 1.0000 1.0000 1.0000 0.4286 1.0000## [7] 1.0000 1.0000 1.0000
transitivity(G4,"global")
## [1] 0.7895
transitivity(G4,"average")
## [1] 0.9365
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The qgraph function clustcoef_auto() will alsoreturn the local clustering coefficient
clustcoef_auto(G4)
## clustWS## 1 1.0000## 2 1.0000## 3 1.0000## 4 1.0000## 5 0.4286## 6 1.0000## 7 1.0000## 8 1.0000## 9 1.0000
Also works for any valid input to qgraph and can returnweighted generalizations of the clustering coefficient (notdiscussed in this course).
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The famous paper of Watts and Strogatz (1998)—alreadycited 23623 times—describes the “small world” principlethat frequently occurs in natural graphs.
I “Six degrees of separation”I High clustering and low average path length
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
mouseover for friend details
Hbased on data from 190 of 203 friendsL
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A graph exhibits a small world structure if it has a muchhigher clustering than a random graph of the samedimensions while still having a low APL.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
# Simulate a graph:set.seed(1)G5 <- watts.strogatz.game(1, 100, 5, 0.05)plot(G5)
1
23
4 56 7
8
9
101112
13
14 1516
17
1819
2021
2223 2425
26
27
2829
30
3132
333435
363738
394041
42
4344
45
46
4748
4950 5152
535455
565758
59
60616263
64 65
66
67
68
6970
717273
74
7576
77 7879
80 8182
83
84
85 86
878889
90
9192
93
9495 96
97
98
99 100
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Function to compute average path length of comparablerandom graph:
APLr <- function(x){if ("qgraph"%in%class(x)) x <- as.igraph(x)if ("igraph"%in%class(x)) x <- get.adjacency(x)
N=nrow(x)p=sum(x/2)/sum(lower.tri(x))
eulers_constant <- .57721566490153l = (log(N)-eulers_constant)/log(p*(N-1)) +.5l
}
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Function to compute clustering of comparable randomgraph:
Cr <- function(x){if ("qgraph"%in%class(x)) x <- as.igraph(x)if ("igraph"%in%class(x)) x <- get.adjacency(x)
N=nrow(x)p=sum(x/2)/sum(lower.tri(x))
t=(p*(N-1)/N)t
}
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Is there a small world?
# Clustering in graph:transitivity(G5)
## [1] 0.5402
# Clustering in random graph:Cr(G5)
## [1] 0.1
# Average path length in graph:average.path.length(G5)
## [1] 2.867
# Average path length in random graph:APLr(G5)
## [1] 2.249
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Small world index:
(transitivity(G5) / Cr(G5)) /(average.path.length(G5) / APLr(G5))
## [1] 4.237
Higher than 3? There is a Small world!
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Alternatively, the qgraph function smallworldness canbe used:
smallworldness(G5)
## smallworldness trans_target## 4.97977 0.54016## averagelength_target trans_rnd_M## 2.86747 0.08413## trans_rnd_lo trans_rnd_up## 0.06950 0.10128## averagelength_rnd_M averagelength_rnd_lo## 2.22391 2.20969## averagelength_rnd_up## 2.23838
Also works for any valid input to qgraph()
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Centrality measures assign numeric values to theimportance of nodes in the graph and answer thequestion “what is the most central node?”.
I DegreeI ClosenessI BetweennessI Eigenvector centrality
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
●
●
● ●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●● ●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
● ●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●●
● ● ●
●●
●●●●
●
●●
●
●
●
●●
●●
●
● ●
●
●●
●
●
●
●
●●●
●●
●●
●
●●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●●
●●
●
●●
●
●●
●●
●
●
●
●●
●●
●
●●●
●
●
●
●
●●● ●
●●
●●●
●
●
●
●
●● ● ●
●
●
●
●
●
●●
●
●
●●
●●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
● ●
●
●
●
●●
●
●
●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●●
● ●●
● ●
●●
● ●
●●
●
●
●●
●
●●
●
● ●●
●
●●
●
●
●●
●
●●
●●
●
●●
● ●●
● ●
●
●
●●●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The degree of node v , CD(v) is simply the number ofedges connected to node v , which we can compute byeither summing over row v or column v of A:
CD(v) =|V |∑i=1
aiv =
|V |∑j=1
avj
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
degree(G)
## [1] 5 1 1 1 6 6 2 2 2 2 2
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The degree distribution, fd , gives the probability that anode in G has degree d :
fd = P (CD(v) = d)
For a given graph the observed degree distribution cansimply be computed by dividing the number of nodes thathave degree d with the total number of nodes:
fd =# of nodes with degree d
|V |
and can easily be represented with an histogram.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
table(degree(G)) / vcount(G)
#### 1 2 5 6## 0.27273 0.45455 0.09091 0.18182
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
library("ggplot2")qplot(degree(G), geom = "histogram")
0
1
2
3
4
5
2 4 6degree(G)
coun
t
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A random graph is a graph in which each edge ispresent with probability p. The degree distribution of arandom graph follows a binomial distribution:
fd =
(|V | − 1
d
)pd(1− p)|V |−1−d
Or Poisson for large graphs:
fd =ρde−ρ
d !
where ρ = |V |p
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Random graphG2 <- erdos.renyi.game(100, 0.3)plot(G2)
1
2
3
4
5
6
7
8
9
10
11
12
1314
15
16
17
18
19
20
21
2223
24
25
2627
2829
3031
32
33
34
35
36
37
38
39
40
4142
43
44
45
46
47
48
49
50
51
52
53
54
55
5657
58
59
60
61
62
63
6465
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
8485
86
8788
89
90
91
92
93 94
95
96
97
98
99
100
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Random graph
qplot(degree(G2), geom = "histogram")
0
5
10
20 25 30 35 40degree(G2)
coun
t
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
High School dating
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Scale-free networks
I Many natural networks do not have a binomial orpoisson degree distribution
I These graphs usually have many nodes with a verylow degree and few nodes with a very high degree
I These are termed scale-free networks, and forthese networks a power-law holds approximately trueat least in part.
fd ∝∼ d−α
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Preferential Attachment
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
G3 <- barabasi.game(50, 1.2, directed=FALSE)plot(G3)
1
2
3
4
5
6
7
8
9
10
1112
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
3637
38
39
40
41
42
4344
45
46
47 48
49
50
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
qplot(degree(G3), geom = "histogram")
0
10
20
30
2.5 5.0 7.5 10.0degree(G3)
coun
t
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Closeness CC(v) defines that a node is central if it is‘close’ to other nodes. This can be computed by takingthe inverse of the sum of all path lengths going from nodev to all other nodes:
CC(v) =1∑|V |
i=1 dist(v , i)
This is only an interesting measure for fully connectedgraphs or components.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
closeness(G)
## [1] 0.06667 0.04167 0.04167 0.04167 0.07143## [6] 0.07143 0.04762 0.04762 0.04762 0.04762## [11] 0.04762
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Betweenness of node v is defined as the sum ofproportions of the number of shortest paths between allpairs of nodes that go through node v :
CB(i) =n∑
i 6=j 6=k∈V
σ(i , j | v)σ(i , j)
Where σ(i , j) is the total number of shortest pathsbetween any two nodes and σ(i , j | v) the amount ofthose paths that go through v .
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
betweenness(G)
## [1] 24.1667 0.0000 0.0000 0.0000 15.0000## [6] 15.0000 0.1667 0.1667 0.1667 0.1667## [11] 0.1667
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Eigenvector centrality states that a node if central if itsneighbors are central, and is recursive:
CEV (v) = α∑{u,v}∈E
CEV (u)
Since A contains only zeroes and ones we can write thisas:
CEV (v) = α
|V |∑i=1
aiv CEV (i)
.We can write this in matrix form:
cEV = αAcEV
, in which:
cEV =
CEV (1)CEV (2)
...CEV (|V |)
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Eigenvector centrality
An eigenvector x of matrix A is defined as follows:
Ax = λx
, in which λ is some scalar and called an eigenvalue.Rearranging the expression on the previous slide leads tothis eigenvalue problem with λ = α−1 and cE I = x
AcEV =1α
cEV
This shows CEV (v) to be the v th element of aneigenvector. Since centrality measures are positive,Perron-Frobenius theorem dictates that this should be theeigenvector corresponding to the largest eigenvalue.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
evcent(G)
## $vector## [1] 0.7384 0.2078 0.2078 0.2078 1.0000 1.0000## [7] 0.5629 0.5629 0.5629 0.5629 0.5629#### $value## [1] 3.553#### $options## $options$bmat## [1] "I"#### $options$n## [1] 11#### $options$which## [1] "LA"#### $options$nev## [1] 1#### $options$tol## [1] 0#### $options$ncv## [1] 0#### $options$ldv## [1] 0#### $options$ishift## [1] 1#### $options$maxiter## [1] 3000#### $options$nb## [1] 1#### $options$mode## [1] 1#### $options$start## [1] 1#### $options$sigma## [1] 0#### $options$sigmai## [1] 0#### $options$info## [1] 0#### $options$iter## [1] 1#### $options$nconv## [1] 1#### $options$numop## [1] 5#### $options$numopb## [1] 0#### $options$numreo## [1] 5
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Centrality recap
Degree How well connected is a node?Closeness How easy is it to reach all other nodes from
a node?Betweenness How well does a node connect other
nodes?Eigenvector Centrality How important are the neighbors
of a node?
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
●
●
● ●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●● ●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
● ●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●●
● ● ●
●●
●●●●
●
●●
●
●
●
●●
●●
●
● ●
●
●●
●
●
●
●
●●●
●●
●●
●
●●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●●
●●
●
●●
●
●●
●●
●
●
●
●●
●●
●
●●●
●
●
●
●
●●● ●
●●
●●●
●
●
●
●
●● ● ●
●
●
●
●
●
●●
●
●
●●
●●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
● ●
●
●
●
●●
●
●
●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●●
● ●●
● ●
●●
● ●
●●
●
●
●●
●
●●
●
● ●●
●
●●
●
●
●●
●
●●
●●
●
●●
● ●●
● ●
●
●
●●●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Degree
●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●● ●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●
●●●●
●●●●
●●●●
●● ●●
●●
●●●●
●●●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●●●●●●●
●●●●
●●
●●●●●●
●●
●●
●●
●●●●●●
●●
●●●●
●●
●●
●●●●
●●●●
●●
●● ●●
●●●●●●
●● ●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●●●
●●●●
●●●●●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●●●●●
●●●●
●●
●●
●● ●●●●
●●●●
●●
●●
●●
●●
●●
●●●●●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●●● ●●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2526
27
28
2930
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
656667
68
69
70
71
72
73
74
75
76
77
78
79
8081
82
83
8485
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103104
105
106
107
108
109
110111
112
113
114
115
116117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132133
134135
136137
138139
140
141
142143
144
145
146
147
148
149
150
151152
153
154155
156
157
158
159
160
161
162
163
164
165
166
167
168
169170
171
172
173174
175
176
177178
179
180
181
182
183
184 185
186
187
188
189
190
191
192
193
194195
196
197
198
199
200
201
202
203
204
205
206
207208
insomnia / difficu...psychomotor agitat...
psychomotor retard...
depressed mood
tachycardia / acce...
often easily distr...
is often touchy or...
anxietynausua
sweating / perspir...Weight loss
difficulty concent...
transient visual, ...
fatigue / fatigue ...
increased appetite
Clinically signifi...
delusions
coma
extreme negativism
wei
ght g
ain
Hypersomnia
feelings of worthl...
Loss of appetite
excessive motor ac...peculiarities of v...
mar
kedl
y di
min
ishe
...
echolalia
echopraxia
recurrent thoughts...
disorganized speech
flight of ideas or...
disorganized behav...
inflated self−este...decreased need for...
more talkative tha...
increase in goal−d...
excessive involvem...
A d
istin
ct p
erio
d ...
motoric immobilitypupillary dilation
mutism
derealization
depersonalization
palpitations
vomiting
fear of one or mor...nystagmus / vertic...
pounding heart
sensations of shor...
abdominal distressThe situations are...
chills or hot flus...
There is evidence ...
fear of losing con...
stupor
elevated blood pre... trembling or shaking
ches
t pai
n or
dis
c...
fear of dying
paresthesias
persistent concern...
worry about the im...
a significant chan...
incoordination
memory impairment ...
The person recogni...
Marked and persist...
rambling flow of t...
slur
red
spee
ch
blurring of vision
tremors
Anxiety about bein...
A distinct period ...
often lies to obta...
often bullies, thr...
often initiates ph...
has used a weapon ... has been physicall...
has been physicall...
has stolen while c...
has forced someone...
has deliberately e...
has deliberately d...
has broken into so...
has stolen items o...
often stays out at...
has run away from ...is often truant fr...
The development of...
often fails to giv...
often has difficul...
often does not see...
ofte
n do
es n
ot fo
l...
often has diff icu...
often avoids, disl...
often loses t hing...is often forgetful...
often fidgets with...
often leaves seat ...
often runs about o...
often has difficul...
is often "on the g...
often talks excess...
often blurts out a...
often has difficul...
often interrupts o...
Excessive anxiety ...
Disturbance of con...
unsteady gait
bradycardia
chills
confusion, seizure...
flat or inappropri...
muscular weakness,...
lowered blood pres...
generalized muscle...
the person experie...
dizziness
lethargy
depressed reflexes
euphoria
Presence of two or...
lacks close friend...
cardiac arrhythmia
piloerection
nervousness
excitement
flushed face
diuresis
gastroint estinal ...
mus
cle
twitc
hing
periods of inexhau...
muscle aches
lacr
imat
ion
or r
hi...
diarrhea
yawning
fever
seiz
ures
Focal neurological...
aphasia
apraxia
agnosia disturbance in exe...
numbness or dimini...
ataxia
dysa
rthr
ia
mus
cle
rigid
ity
hyperacusis
Confusion about pe...
The
per
son
finds
i...
mind going blank
mus
cle
tens
ion
odd beliefs or mag...unusual perceptual...
decreased heart rate
vivid, unpleasant ...
suspiciousness or ...
The traumatic even...
transient, stress ...
Marked avoidance o...
ideas of reference
A change in cognit...
development of a p...
Recent ingestion o...
autonomic hyperact...
increased hand tre...
grand mal seizures
feelings of hopele...
dissociative amnesia
behavior or appear...
frantic efforts to...
a pa
ttern
of u
nsta
...
impulsivity in at ...
recurrent suicidal...
affective instabil...
chronic fee lings ...
failure to conform...deceitfulness
Per
sist
ent a
void
an...
often loses temperoften argues with ...
ofte
n de
liber
atel
y...
often blames other...
is often angry and...
is often spiteful ...
hypervigilance
exaggerated startl...
impulsivity or fai...
reckless disregard...
consistent irrespo...
lack of remorse
often actively def... Dysphoric mood
drowsiness
neither desires no...
has little, if any...
takes pleasure in ...
Pupillary constric...
catalepsy
almost always choo...
appears indifferen...
conjunctival injec...
dry
mou
th
Relative unrespons...
No detailed dream ...
Rec
urre
nt e
piso
des.
..
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Closeness
●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●● ●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●
●●●●
●●●●
●●●●
●● ●●
●●
●●●●
●●●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●●●●●●●
●●●●
●●
●●●●●●
●●
●●
●●
●●●●●●
●●
●●●●
●●
●●
●●●●
●●●●
●●
●● ●●
●●●●●●
●● ●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●●●
●●●●
●●●●●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●●●●●
●●●●
●●
●●
●● ●●●●
●●●●
●●
●●
●●
●●
●●
●●●●●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●●● ●●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2526
27
28
2930
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
656667
68
69
70
71
72
73
74
75
76
77
78
79
8081
82
83
8485
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103104
105
106
107
108
109
110111
112
113
114
115
116117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132133
134135
136137
138139
140
141
142143
144
145
146
147
148
149
150
151152
153
154155
156
157
158
159
160
161
162
163
164
165
166
167
168
169170
171
172
173174
175
176
177178
179
180
181
182
183
184 185
186
187
188
189
190
191
192
193
194195
196
197
198
199
200
201
202
203
204
205
206
207208
insomnia / difficu...psychomotor agitat...
increased appetitedepressed mood
psychomotor retard...
nausua
transient visual, ...
difficulty concent...
fatigue / fatigue ...
Weight loss
weight gain
sweating / perspir...is often touchy or...
anxiety
tach
ycar
dia
/ acc
e...
vomiting
often easily distr...
pupillary dilation
delusions
disorganized speech
diso
rgan
ized
beh
av...
extreme negativism
Hypersomnia
feel
ings
of w
orth
l...
echolalia
echopraxia
mutism
com
a
palpitations
incoordination
mus
cle
ache
s
diarrheayawning
fever chills
mus
cula
r w
eakn
ess,
...
tremors
excitement
flush
ed fa
ce
stupor
leth
argy
ataxia
unexpected travel ...
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Betweenness
●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●● ●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●
●●●●
●●●●
●●●●
●● ●●
●●
●●●●
●●●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●●●●●●●
●●●●
●●
●●●●●●
●●
●●
●●
●●●●●●
●●
●●●●
●●
●●
●●●●
●●●●
●●
●● ●●
●●●●●●
●● ●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●●●
●●●●
●●●●●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●●●●●
●●●●
●●
●●
●● ●●●●
●●●●
●●
●●
●●
●●
●●
●●●●●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●●● ●●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2526
27
28
2930
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
656667
68
69
70
71
72
73
74
75
76
77
78
79
8081
82
83
8485
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103104
105
106
107
108
109
110111
112
113
114
115
116117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132133
134135
136137
138139
140
141
142143
144
145
146
147
148
149
150
151152
153
154155
156
157
158
159
160
161
162
163
164
165
166
167
168
169170
171
172
173174
175
176
177178
179
180
181
182
183
184 185
186
187
188
189
190
191
192
193
194195
196
197
198
199
200
201
202
203
204
205
206
207208
is often touchy or...often easily distr...
depressed mood
insomnia / difficu...
anxietypsychomotor agitat...
tachycardia / acce...
psychomotor retard...
naus
ua
sweating / perspir...flat or inappropri...
increased appetitefear of one or mor...
Clinically signifi...
transient visual, ...
rambling flow of t...
pupillary dilation
difficulty concent...
memory impairment ...
Weight loss
coma
disorganized speech
disorganized behav...
extreme negativism
fatigue / fatigue ...
derealization
depersonalization
delusions
Disturbance of con...
vom
iting
weight gainvivid, unpleasant ...
Con
fusi
on a
bout
pe.
..
lacks close friend...
There is evidence ...
Excessive anxiety ...
palpitations
elevated blood pre...
the person experie...
stupor
nystagmus / vertic...
incoordination
blurring of vision
tremors
odd beliefs or mag...
unusual perceptual...
excessive motor ac...
peculiarities of v... echolalia
echopraxia
Hypersomnia
slurred speech
Focal neurological...
aphasia
apraxia
agnosia disturbance in exe...
feelings of worthl...
Loss of appetite
unsteady gait
markedly diminishe...recurrent thoughts...often lies to obta...
often bullies, thr...
often initiates ph...
has used a weapon ...
has been physicall... has been physicall...
has stolen while c... has forced someone...
has deliberately e...has deliberately d...
has broken into so...
has stolen items o...
often stays out at...
has run away from ...
is often truant fr...
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Eigenvector Centrality
●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●● ●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●
●●●●
●●●●
●●●●
●● ●●
●●
●●●●
●●●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●●●●●●●
●●●●
●●
●●●●●●
●●
●●
●●
●●●●●●
●●
●●●●
●●
●●
●●●●
●●●●
●●
●● ●●
●●●●●●
●● ●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●●●
●●●●
●●●●●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●●●●●
●●●●
●●
●●
●● ●●●●
●●●●
●●
●●
●●
●●
●●
●●●●●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●●● ●●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2526
27
28
2930
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
656667
68
69
70
71
72
73
74
75
76
77
78
79
8081
82
83
8485
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103104
105
106
107
108
109
110111
112
113
114
115
116117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132133
134135
136137
138139
140
141
142143
144
145
146
147
148
149
150
151152
153
154155
156
157
158
159
160
161
162
163
164
165
166
167
168
169170
171
172
173174
175
176
177178
179
180
181
182
183
184 185
186
187
188
189
190
191
192
193
194195
196
197
198
199
200
201
202
203
204
205
206
207208
psychomotor agitat...insomnia / difficu...
psychomotor retard...Weight loss
depressed moodtransient visual, ...
difficulty concent...
fatigue / fatigue ...
increased appetite
often easily distr...
wei
ght g
ain
feelings of worthl...
Loss of appetite
Hypersomnia
mar
kedl
y di
min
ishe
...
recurrent thoughts...
decreased need for...
excessive involvem...
more talkative tha...
delusions
echopraxia
echolalia
mot
oric
imm
obili
tymutismA distinct period ...
Presence of two or...
tachycardia / acce...
anxietynausua
sweating / perspir...
is often touchy or...
Clinically signifi...
pupillary dilation
vomitingcoma
feel
ings
of h
opel
e...
stupor
decreased heart rate
elev
ated
blo
od p
re...
vivid, unpleasant ...
bradycardia
chills
confusion, seizure...
muscular weakness,...lowered blood pres...
Dysphoric mood
Exc
essi
ve a
nxie
ty ..
.
mind going blank
The person finds i...
muscle tension
incoordination
blurring of vision
tremors
catalepsy
palpitations
grand mal seizures
increased hand tre...
auto
nom
ic h
yper
act..
.
memory impairment ...
nystagmus / vertic...
ram
blin
g flo
w o
f t...
the person experie...
slurred speech
exci
tem
ent
flushed face
card
iac
arrh
ythm
ia
nervousness
diuresis
mus
cle
twitc
hing
gastroint estinal ...
periods of inexhau...
depersonalization
derealizationyawning
piloerection
fever
lacrimation or rhi...
diarrhea
muscle aches
unsteady gait
hypervigilancePersistent avoidan...
exaggerated startl...
euphoria
leth
argy
generalized muscle...
depressed reflexes
dizziness
fear of one or mor...
paresthesias
chest pain or disc...
fear
of d
ying
The situations are...
sensations of shor...
abdominal distress
fear of losing con...
persistent concern...
pounding heart
a significant chan...
chills or hot flus...
worry about the im...
trembling or shaking
The
per
son
reco
gni..
.
often bullies, thr...
has been physicall... often lies to obta...
has used a weapon ...
has forced someone...
has deliberately e...
has been physicall...
has deliberately d...
has broken into so...
has stolen items o...
often initiates ph...
has stolen while c...
has run away from ...
often stays out at...
is often truant fr...
The
dev
elop
men
t of..
.
Marked and persist...
Anxiety about bein...
aphasia
apraxia
agnosia
disturbance in exe...
Focal neurological...
flat or inappropri...
often avoids, disl... is often forgetful...
often has difficul...
often does not see...
often fails to giv...often has diff icu...
often runs about o...
often loses t hing...
often fidgets with...
often leaves seat ...
often interrupts o...
often does not fol...
is often "on the g...
often has difficul...
ofte
n ha
s di
fficu
l...
often talks excess...
often blurts out a...
conjunctival injec...
dry mouth
There is evidence ...
numbness or dimini...muscle rigidity
seizures
hype
racu
sis
dysarthria ataxia
Disturbance of con...
Recent ingestion o...
unusual perceptual...
odd beliefs or mag...
drowsiness
Pup
illar
y co
nstr
ic...
Marked avoidance o...
The traumatic even...
dissociative amnesia
A change in cognit...
development of a p...
impulsivity or fai...
reckless disregard...
deceitfulness
failure to conform...
lack of remorse
consistent irrespo...
Confusion about pe...
frantic efforts to...
recurrent suicidal...
impulsivity in at ...
affective instabil...
transient, stress ...
a pattern of unsta...
chronic fee lings ...
often argues with ...
often deliberately...
often blames other...
often loses temper
is often angry and...
often actively def...
is often spiteful ...
No detailed dream ...
Recurrent episodes...
Relative unrespons...
lacks close friend...
behavior or appear...
ideas of reference
suspiciousness or ...
On awakening from ...
almost always choo...has little, if any...
neither desires no...
appears indifferen...
takes pleasure in ...
unexpected travel ...
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
How to win VAX
I Delete nodes that haveI High centralityI Low clustering
I Reduce small-worldnessI Increase the average shortest path length
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1 2
3
4
5
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A weighted graph is a network in which the strength ofconnections can vary. For a graph of n nodes the networkstructure is defined by the |V | × |V | adjacency matrix Asuch that:
aij =
{1 if there is an edge from node i to node j0 otherwise
.Its weights are defined by the |V | × |V | weights matrix Wsuch that:
wij
{= 0 if aij = 0∈ R otherwise
.The diagonals of both A and W are set to zero.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A graph is undirected only if A and W are symmetric:
A = A>
W = W>
A graph is unweighted only if A is equal to W multipliedby some scalar c:
A = cW
Note that both these cases do not imply that a graph isundirected or unweighted.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Finally, we define the length of an edge from node i tonode j as the inverse of the absolute weight:
lij =
{1|wij | if i 6= j
0 if i = j
Because the denumerator is always positive, we will takethe limit in the case of a weight of 0:
limwij→0
1|wij |
=∞
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
I In unweighted graphs network descriptives are welldefined
I In weighted graphs this is less the case. There is alot of debate on whether the structure (adjacencymatrix) or the weights (weights matrix) are the mostimportant
I Classically only the weights matrix is analyzedI Opsahl, Agneessens, and Skvoretz (2010) proposes
a set of centrality measures with a tuning parameter,α, to manually assign importance to both matrices
I α of 0 only regards the structure, and α of 1 onlyregards the weights
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
A =
0 1 10 0 10 0 0
W =
0 1 20 0 30 0 0
L =
0 1 1/2∞ 0 1/3∞ ∞ 0
1
2
3
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
The shortest path length between nodes i and j , d(i , j) isdefined in an unweighted graph as the minimum numberof steps you need to take from node i to node j :
d(i , j) = min(aih + . . .+ ahj
)which can be obtained through Dijkstra’s algorithm(Dijkstra, 1959) with weights fixed to 1.Similarly, for weighted graphs the shortest path length isdefined as the minimum number of distance needed tocross on the graph to reach node i from node j :
d(i , j) = min(
1|wih|
+ . . .+1|whj |
)= min
(lih + . . .+ lhj
)
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Opsahl et al. (2010) propose the following combination:
d(i , j) = min(
lαih + . . .+ lαhj
)Which generalizes to the unweighted form near α = 0and the weighted form if α = 1.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Of node i the in-degree k (in)i is the number of incoming
edges and the out-degree k (out)i the number of outgoing
edges:
k (in)i =
n∑j=1
aji
k (out)i =
n∑j=1
aij
In weighted graphs often the node strengths s(in)i and
s(out)i are used:
s(in)i =
n∑j=1
wji
s(out)i =
n∑j=1
wij
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Opsahl et al. (2010) propose the following combination:
C(in)D (i) = k (in)
i ×
(s(in)
i
k (in)i
)α
C(out)D (i) = k (out)
i ×
(s(out)
i
k (out)i
)α
Note that for both:
CD(i) =
{ki if α = 0si if α = 1
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Closeness is defined as the inverse of the total shortestdistance from node i to all other nodes in the graph,which is only defined for connected clusters:
CC(i) =
n∑j=1
d(i , j)
−1
Here the α parameter is already used in computing theshortest path lengths.
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Betweenness of node i is defined as the sum ofproportions of the number of shortest paths between allpairs of nodes that go through node i :
CB(i) =n∑
i 6=j 6=k
gjk (i)gjk
Where gjk is the total number of shortest paths betweenany two nodes and gjk (i) the amount of those paths thatgo through i .
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1 2
3
4
5
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1 2
3
4
5
Dout
Din
Din
DinC
B
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Dout
DoutDin
Din
C
C
B
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Dout
Dout
Dout
Din
Din
Din
Din
Din
C
B
B
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
What is the most central node?
1
2
3
4
5
6
7
8
9
10
11
Dout
Din
C
B
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
Literature on Blackboard:
I Node centrality in weighted networks: Generalizingdegree and shortest paths
I Collective dynamics of ‘small-world’ networksI The Small World of PsychopathologyI State of the aRt personality research: A tutorial on
network analysis of personality data in R
Graph descriptives
Sacha Epskamp
Recap
Introduction
ConnectivityShortest Path Length
Diameter and Density
ClusteringLocal Clustering
Global Clustering
Small-worldness
CentralityDegree
Degree distribution
Closeness
Betweenness
Eigenvector centrality
Weighted andDirected networksShortest Path length
Centrality
References
References I
Dijkstra, E. (1959). A note on two problems in connexionwith graphs. Numerische mathematik, 1(1),269–271.
Opsahl, T., Agneessens, F., & Skvoretz, J. (2010). Nodecentrality in weighted networks: Generalizingdegree and shortest paths. Social Networks,32(3), 245–251.
Watts, D., & Strogatz, S. (1998). Collective dynamics ofsmall-world networks. Nature, 393(6684),440–442.