淺談社會網絡分析

Introduction to Social Network Analysis

新手向 BY 3.0 TW

文月 (Meng-Ying Tsai)

聽說有八盤之類的奇妙綽號
台大兔子系四年級
偶而會在台大開源社出沒
❤: 喝淺焙咖啡、唱日卡、嚐甜食

一年365天歡迎餵食，請多指教ヽ(●´∀`●)ﾉ

Vertex

Edge

Vertex Set V(G)

Edge Set E(G)

Graph G = (V, E)

V = {1, 2, 3, 4}

E = {{1, 2}, {1, 3}, {1, 4},

{2, 3}, {3, 4}}

1

2

3

4

{1,2}

{1,3}

{2,3}

{1,4}

{3,4}

1

2

3

4

{1,2}

{1,3}

{2,3}

{1,4}

{3,4}

	1	2	3	4
1	0	1	1	1
2	1	0	1	0
3	1	1	0	1
4	1	0	1	0

Basic Construct

Cohesion

Centrality

Cluster

* 接下來皆以無向圖作舉例！

Basic Construct

Cohesion

Density
Distance
Connectivity

Basic Construct

Density

~~誰會贏?~~ 誰 density 高?

Density

Distance

Connectivity

Basic Construct

Density

D =

|E|

|V| * (|V|-1)

2

1

# of edge

# of total possible edge

=

100%

50%

Density

Distance

Connectivity

Basic Construct

Distance

Six Degrees of Separation

3.57 Degrees of Separation

Density

Distance

Connectivity

Basic Construct

Distance

geodesic distance d(x,y): shortest path from x to y

x

y

Density

Distance

Connectivity

Basic Construct

Distance

diameter: largest geodesic distance between any pair of nodes

x

y

Density

Distance

Connectivity

Basic Construct

Connectivity

不能沒有你QQ

Density

Distance

Connectivity

Basic Construct

Connectivity

Density

Distance

Connectivity

point connectivity: the min # of nodes that must be removed to disconnect 2 nodes

x

y

1

2

Basic Construct

Connectivity

Density

Distance

Connectivity

太過依賴單一個頂點，
網絡就顯得特別脆弱QQ

Basic Construct

Degree
Closeness
Betweenness
Eigenvector

Centrality

常見有 4 種

Basic Construct

Centrality

Degree Centrality: 該 node 的連線越多，centrality 越高

覺得邊QAQ"

Basic Construct

Centrality

Closeness Centrality: geodesic distance 加總取倒數，distance越短centrality越高。

覺得邊

QAQ"

Basic Construct

Centrality

Betweenness Centrality: 越常座落在別人的 geodesic path，centrality 越高。

Basic Construct

Centrality

Eigenvector Centrality: 基本概念為 degree centrality，跟越重要的 node 連，算分越高。

認識世界知名人物

認識默默無名小生物

I'm here

>

Pigeon

Degree
Closeness
Betweenness
Eigenvector

Basic Construct

Clique
K-core

Cluster

Basic Construct

Cluster

n-Clique: a maximal subgraph in which every pair of vertices is connected by a path of length n or less

n-Clique
K-core

n=1

n=2

maximal subgraph：再加入任一頂點就無法維持其性質
點跟點之間的距離 ≤ n

Basic Construct

Cluster

n-Clique: a maximal subgraph in which every pair of vertices is connected by a path of length n or less

條件嚴格，形成 clique 難
每個 clique 都被視為一樣重要

n-Clique
K-core

n=1

n=2

Basic Construct

Cluster

K-core: a subgroup is defined as k-core when member have directed ties to at least K other vertices.

n-Clique
K-core

K = 3

K = 2

K = 1

n-Clique
K-core

K = 3

K = 2

K = 1

~~其實標題什麼的，都是胡扯~~

~~暗黑破壞神~~

~~暗黑復仇者~~

~~Dark Web~~

暗黑網絡

Dark/Covert network!

Cases

Dark Network

Krebs V.(2001).Mapping Networks of Terrorist Cells. Connections, 24(3)

911 hijacking data

trusted prior contacts

鬆散
連同組的都要 2 steps way 才能接觸
避免其中有人被揪出來，整個網絡被連根拔起

Cases

Dark Network

Krebs V.(2001).Mapping Networks of Terrorist Cells. Connections, 24(3)

911 hijakcing data

trusted prior contacts + ties

太鬆散無法做事QQ
讓溝通的順利的meeting!
沒有需要時，立即進入冬眠狀態

Cases

Dark Network

Hughes C.E., Chalmers J., Bright D.A., McFadden M. (2017) Social network analysis of Australian poly-drug trafficking networks: How do drug traffickers manage multiple illicit drugs? Social Networks, 51C

Poly-drug network

藥物買賣(X) 毒品交易(O)

顏色標示不同毒品種類

大麻 MDMA 甲基安非他命

顏色表示身份

manager resource provider

worker wholesale supplier

Cases

Dark Network

Hughes C.E., Chalmers J., Bright D.A., McFadden M. (2017) Social network analysis of Australian poly-drug trafficking networks: How do drug traffickers manage multiple illicit drugs? Social Networks, 51C

Poly-drug network

顏色表示身份

manager resource provider

worker wholesale supplier

什麼毒品都賣~~，都可以賣~~
沒有鬆散、去中心的特質

Cases

Dark Network

Hughes C.E., Chalmers J., Bright D.A., McFadden M. (2017) Social network analysis of Australian poly-drug trafficking networks: How do drug traffickers manage multiple illicit drugs? Social Networks, 51C