Suche senden
Hochladen
Social network analysis part ii
•
19 gefällt mir
•
6,601 views
THomas Plotkowiak
Folgen
An overview of
Weniger lesen
Mehr lesen
Bildung
Technologie
Business
Melden
Teilen
Melden
Teilen
1 von 129
Jetzt herunterladen
Downloaden Sie, um offline zu lesen
Empfohlen
Social Network Analysis Workshop
Social Network Analysis Workshop
Data Works MD
Social Network Analysis (SNA) 2018
Social Network Analysis (SNA) 2018
Arsalan Khan
The Basics of Social Network Analysis
The Basics of Social Network Analysis
Rory Sie
Community Detection in Social Networks: A Brief Overview
Community Detection in Social Networks: A Brief Overview
Satyaki Sikdar
Exploratory social network analysis with pajek
Exploratory social network analysis with pajek
THomas Plotkowiak
Social Network Analysis
Social Network Analysis
Fred Stutzman
Social Network Analysis: What It Is, Why We Should Care, and What We Can Lear...
Social Network Analysis: What It Is, Why We Should Care, and What We Can Lear...
Xiaohan Zeng
Community detection algorithms
Community detection algorithms
Alireza Andalib
Empfohlen
Social Network Analysis Workshop
Social Network Analysis Workshop
Data Works MD
Social Network Analysis (SNA) 2018
Social Network Analysis (SNA) 2018
Arsalan Khan
The Basics of Social Network Analysis
The Basics of Social Network Analysis
Rory Sie
Community Detection in Social Networks: A Brief Overview
Community Detection in Social Networks: A Brief Overview
Satyaki Sikdar
Exploratory social network analysis with pajek
Exploratory social network analysis with pajek
THomas Plotkowiak
Social Network Analysis
Social Network Analysis
Fred Stutzman
Social Network Analysis: What It Is, Why We Should Care, and What We Can Lear...
Social Network Analysis: What It Is, Why We Should Care, and What We Can Lear...
Xiaohan Zeng
Community detection algorithms
Community detection algorithms
Alireza Andalib
06 Community Detection
06 Community Detection
Duke Network Analysis Center
Community detection in graphs
Community detection in graphs
Nicola Barbieri
Social Network Analysis
Social Network Analysis
Giorgos Cheliotis
Community detection in social networks
Community detection in social networks
Francisco Restivo
Network centrality measures and their effectiveness
Network centrality measures and their effectiveness
emapesce
3 Centrality
3 Centrality
Maksim Tsvetovat
Social Media Mining - Chapter 8 (Influence and Homophily)
Social Media Mining - Chapter 8 (Influence and Homophily)
SocialMediaMining
Network measures used in social network analysis
Network measures used in social network analysis
Dragan Gasevic
Link prediction
Link prediction
Carlos Castillo (ChaTo)
Social network analysis course 2010 - 2011
Social network analysis course 2010 - 2011
guillaume ereteo
4 Cliques Clusters
4 Cliques Clusters
Maksim Tsvetovat
Community detection
Community detection
Scott Pauls
GraphSage vs Pinsage #InsideArangoDB
GraphSage vs Pinsage #InsideArangoDB
ArangoDB Database
Social Media Mining - Chapter 7 (Information Diffusion)
Social Media Mining - Chapter 7 (Information Diffusion)
SocialMediaMining
4. social network analysis
4. social network analysis
Lokesh Ramaswamy
Community Detection in Social Media
Community Detection in Social Media
Symeon Papadopoulos
Social Network Analysis Introduction including Data Structure Graph overview.
Social Network Analysis Introduction including Data Structure Graph overview.
Doug Needham
Introduction to Social Network Analysis
Introduction to Social Network Analysis
Toronto Metropolitan University
07 Whole Network Descriptive Statistics
07 Whole Network Descriptive Statistics
Duke Network Analysis Center
Social Network Analysis Using Gephi
Social Network Analysis Using Gephi
Goa App
Social network analysis intro part I
Social network analysis intro part I
THomas Plotkowiak
Exploratory facebook social network analysis with gephi
Exploratory facebook social network analysis with gephi
THomas Plotkowiak
Weitere ähnliche Inhalte
Was ist angesagt?
06 Community Detection
06 Community Detection
Duke Network Analysis Center
Community detection in graphs
Community detection in graphs
Nicola Barbieri
Social Network Analysis
Social Network Analysis
Giorgos Cheliotis
Community detection in social networks
Community detection in social networks
Francisco Restivo
Network centrality measures and their effectiveness
Network centrality measures and their effectiveness
emapesce
3 Centrality
3 Centrality
Maksim Tsvetovat
Social Media Mining - Chapter 8 (Influence and Homophily)
Social Media Mining - Chapter 8 (Influence and Homophily)
SocialMediaMining
Network measures used in social network analysis
Network measures used in social network analysis
Dragan Gasevic
Link prediction
Link prediction
Carlos Castillo (ChaTo)
Social network analysis course 2010 - 2011
Social network analysis course 2010 - 2011
guillaume ereteo
4 Cliques Clusters
4 Cliques Clusters
Maksim Tsvetovat
Community detection
Community detection
Scott Pauls
GraphSage vs Pinsage #InsideArangoDB
GraphSage vs Pinsage #InsideArangoDB
ArangoDB Database
Social Media Mining - Chapter 7 (Information Diffusion)
Social Media Mining - Chapter 7 (Information Diffusion)
SocialMediaMining
4. social network analysis
4. social network analysis
Lokesh Ramaswamy
Community Detection in Social Media
Community Detection in Social Media
Symeon Papadopoulos
Social Network Analysis Introduction including Data Structure Graph overview.
Social Network Analysis Introduction including Data Structure Graph overview.
Doug Needham
Introduction to Social Network Analysis
Introduction to Social Network Analysis
Toronto Metropolitan University
07 Whole Network Descriptive Statistics
07 Whole Network Descriptive Statistics
Duke Network Analysis Center
Social Network Analysis Using Gephi
Social Network Analysis Using Gephi
Goa App
Was ist angesagt?
(20)
06 Community Detection
06 Community Detection
Community detection in graphs
Community detection in graphs
Social Network Analysis
Social Network Analysis
Community detection in social networks
Community detection in social networks
Network centrality measures and their effectiveness
Network centrality measures and their effectiveness
3 Centrality
3 Centrality
Social Media Mining - Chapter 8 (Influence and Homophily)
Social Media Mining - Chapter 8 (Influence and Homophily)
Network measures used in social network analysis
Network measures used in social network analysis
Link prediction
Link prediction
Social network analysis course 2010 - 2011
Social network analysis course 2010 - 2011
4 Cliques Clusters
4 Cliques Clusters
Community detection
Community detection
GraphSage vs Pinsage #InsideArangoDB
GraphSage vs Pinsage #InsideArangoDB
Social Media Mining - Chapter 7 (Information Diffusion)
Social Media Mining - Chapter 7 (Information Diffusion)
4. social network analysis
4. social network analysis
Community Detection in Social Media
Community Detection in Social Media
Social Network Analysis Introduction including Data Structure Graph overview.
Social Network Analysis Introduction including Data Structure Graph overview.
Introduction to Social Network Analysis
Introduction to Social Network Analysis
07 Whole Network Descriptive Statistics
07 Whole Network Descriptive Statistics
Social Network Analysis Using Gephi
Social Network Analysis Using Gephi
Andere mochten auch
Social network analysis intro part I
Social network analysis intro part I
THomas Plotkowiak
Exploratory facebook social network analysis with gephi
Exploratory facebook social network analysis with gephi
THomas Plotkowiak
Social tv
Social tv
THomas Plotkowiak
Memes
Memes
THomas Plotkowiak
Information Visualization for Social Network Analysis,
Information Visualization for Social Network Analysis,
University of Maryland
Visualizacion de Redes en Curso gratis de Grafos
Visualizacion de Redes en Curso gratis de Grafos
Educagratis
Finding All Maximal Cliques in Very Large Social Networks
Finding All Maximal Cliques in Very Large Social Networks
Antonio Maccioni
Accordion - VLDB 2014
Accordion - VLDB 2014
Marco Serafini
Ongoing Research on Politics & Twitter - CSS Colloquium
Ongoing Research on Politics & Twitter - CSS Colloquium
Talha OZ
Terrified
Terrified
Talha OZ
XXL Graph Algorithms__HadoopSummit2010
XXL Graph Algorithms__HadoopSummit2010
Yahoo Developer Network
Importance
Importance
Naoki Masuda
Collective Stress in the Digital Age
Collective Stress in the Digital Age
Talha OZ
Network Analysis with networkX : Fundamentals of network theory-1
Network Analysis with networkX : Fundamentals of network theory-1
Kyunghoon Kim
CLIQUE Automatic subspace clustering of high dimensional data for data mining...
CLIQUE Automatic subspace clustering of high dimensional data for data mining...
Raed Aldahdooh
Análise de redes sociais, parte 3
Análise de redes sociais, parte 3
Francisco Restivo
Tecnologia e Empreendedorismo
Tecnologia e Empreendedorismo
Francisco Restivo
Services marketing- Albaraka bank pakistan
Services marketing- Albaraka bank pakistan
Ashiq Hussain
Vertex neighborhoods, low conductance cuts, and good seeds for local communit...
Vertex neighborhoods, low conductance cuts, and good seeds for local communit...
David Gleich
Al baraka pakistan
Al baraka pakistan
Saad Amin
Andere mochten auch
(20)
Social network analysis intro part I
Social network analysis intro part I
Exploratory facebook social network analysis with gephi
Exploratory facebook social network analysis with gephi
Social tv
Social tv
Memes
Memes
Information Visualization for Social Network Analysis,
Information Visualization for Social Network Analysis,
Visualizacion de Redes en Curso gratis de Grafos
Visualizacion de Redes en Curso gratis de Grafos
Finding All Maximal Cliques in Very Large Social Networks
Finding All Maximal Cliques in Very Large Social Networks
Accordion - VLDB 2014
Accordion - VLDB 2014
Ongoing Research on Politics & Twitter - CSS Colloquium
Ongoing Research on Politics & Twitter - CSS Colloquium
Terrified
Terrified
XXL Graph Algorithms__HadoopSummit2010
XXL Graph Algorithms__HadoopSummit2010
Importance
Importance
Collective Stress in the Digital Age
Collective Stress in the Digital Age
Network Analysis with networkX : Fundamentals of network theory-1
Network Analysis with networkX : Fundamentals of network theory-1
CLIQUE Automatic subspace clustering of high dimensional data for data mining...
CLIQUE Automatic subspace clustering of high dimensional data for data mining...
Análise de redes sociais, parte 3
Análise de redes sociais, parte 3
Tecnologia e Empreendedorismo
Tecnologia e Empreendedorismo
Services marketing- Albaraka bank pakistan
Services marketing- Albaraka bank pakistan
Vertex neighborhoods, low conductance cuts, and good seeds for local communit...
Vertex neighborhoods, low conductance cuts, and good seeds for local communit...
Al baraka pakistan
Al baraka pakistan
Ähnlich wie Social network analysis part ii
02 Introduction to Social Networks and Health: Key Concepts and Overview
02 Introduction to Social Networks and Health: Key Concepts and Overview
Duke Network Analysis Center
00 Introduction to SN&H: Key Concepts and Overview
00 Introduction to SN&H: Key Concepts and Overview
Duke Network Analysis Center
The effect of social welfare
The effect of social welfare
csandit
The Effect of Social Welfare System Based on the Complex Network
The Effect of Social Welfare System Based on the Complex Network
csandit
01 Introduction to Networks Methods and Measures
01 Introduction to Networks Methods and Measures
dnac
01 Introduction to Networks Methods and Measures (2016)
01 Introduction to Networks Methods and Measures (2016)
Duke Network Analysis Center
What isa border_kings
What isa border_kings
INRIA - ENS Lyon
Fbk Seminar Michela Ferron
Fbk Seminar Michela Ferron
Bruno Kessler Foundation
MODELING SOCIAL GAUSS-MARKOV MOBILITY FOR OPPORTUNISTIC NETWORK
MODELING SOCIAL GAUSS-MARKOV MOBILITY FOR OPPORTUNISTIC NETWORK
csandit
Legal Analytics Course - Class 11 - Network Analysis and Law - Professors Dan...
Legal Analytics Course - Class 11 - Network Analysis and Law - Professors Dan...
Daniel Katz
Adaptive network models of socio-cultural dynamics
Adaptive network models of socio-cultural dynamics
Hiroki Sayama
Temporal dynamics of human behavior in social networks (i)
Temporal dynamics of human behavior in social networks (i)
Esteban Moro
Node similarity
Node similarity
SURAJ NAYAK
01 Network Data Collection (2017)
01 Network Data Collection (2017)
Duke Network Analysis Center
SSRI_pt1.ppt
SSRI_pt1.ppt
9260SahilPatil
Social and economical networks from (big-)data - Esteban Moro II
Social and economical networks from (big-)data - Esteban Moro II
Lake Como School of Advanced Studies
cuhk-fb-mi-talk.pdf
cuhk-fb-mi-talk.pdf
Laks Lakshmanan
Temporal networks - Alain Barrat
Temporal networks - Alain Barrat
Lake Como School of Advanced Studies
Actor Network Theory - Modernity to Postmodernity
Actor Network Theory - Modernity to Postmodernity
Craig Hammond
Mathematics and Social Networks
Mathematics and Social Networks
Mason Porter
Ähnlich wie Social network analysis part ii
(20)
02 Introduction to Social Networks and Health: Key Concepts and Overview
02 Introduction to Social Networks and Health: Key Concepts and Overview
00 Introduction to SN&H: Key Concepts and Overview
00 Introduction to SN&H: Key Concepts and Overview
The effect of social welfare
The effect of social welfare
The Effect of Social Welfare System Based on the Complex Network
The Effect of Social Welfare System Based on the Complex Network
01 Introduction to Networks Methods and Measures
01 Introduction to Networks Methods and Measures
01 Introduction to Networks Methods and Measures (2016)
01 Introduction to Networks Methods and Measures (2016)
What isa border_kings
What isa border_kings
Fbk Seminar Michela Ferron
Fbk Seminar Michela Ferron
MODELING SOCIAL GAUSS-MARKOV MOBILITY FOR OPPORTUNISTIC NETWORK
MODELING SOCIAL GAUSS-MARKOV MOBILITY FOR OPPORTUNISTIC NETWORK
Legal Analytics Course - Class 11 - Network Analysis and Law - Professors Dan...
Legal Analytics Course - Class 11 - Network Analysis and Law - Professors Dan...
Adaptive network models of socio-cultural dynamics
Adaptive network models of socio-cultural dynamics
Temporal dynamics of human behavior in social networks (i)
Temporal dynamics of human behavior in social networks (i)
Node similarity
Node similarity
01 Network Data Collection (2017)
01 Network Data Collection (2017)
SSRI_pt1.ppt
SSRI_pt1.ppt
Social and economical networks from (big-)data - Esteban Moro II
Social and economical networks from (big-)data - Esteban Moro II
cuhk-fb-mi-talk.pdf
cuhk-fb-mi-talk.pdf
Temporal networks - Alain Barrat
Temporal networks - Alain Barrat
Actor Network Theory - Modernity to Postmodernity
Actor Network Theory - Modernity to Postmodernity
Mathematics and Social Networks
Mathematics and Social Networks
Kürzlich hochgeladen
Seal of Good Local Governance (SGLG) 2024Final.pptx
Seal of Good Local Governance (SGLG) 2024Final.pptx
negromaestrong
Z Score,T Score, Percential Rank and Box Plot Graph
Z Score,T Score, Percential Rank and Box Plot Graph
Thiyagu K
psychiatric nursing HISTORY COLLECTION .docx
psychiatric nursing HISTORY COLLECTION .docx
PoojaSen20
Ecological Succession. ( ECOSYSTEM, B. Pharmacy, 1st Year, Sem-II, Environmen...
Ecological Succession. ( ECOSYSTEM, B. Pharmacy, 1st Year, Sem-II, Environmen...
Shubhangi Sonawane
Gardella_Mateo_IntellectualProperty.pdf.
Gardella_Mateo_IntellectualProperty.pdf.
MateoGardella
Unit-V; Pricing (Pharma Marketing Management).pptx
Unit-V; Pricing (Pharma Marketing Management).pptx
VishalSingh1417
Nutritional Needs Presentation - HLTH 104
Nutritional Needs Presentation - HLTH 104
misteraugie
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdf
Admir Softic
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
kauryashika82
fourth grading exam for kindergarten in writing
fourth grading exam for kindergarten in writing
TeacherCyreneCayanan
Measures of Dispersion and Variability: Range, QD, AD and SD
Measures of Dispersion and Variability: Range, QD, AD and SD
Thiyagu K
SOCIAL AND HISTORICAL CONTEXT - LFTVD.pptx
SOCIAL AND HISTORICAL CONTEXT - LFTVD.pptx
iammrhaywood
An Overview of Mutual Funds Bcom Project.pdf
An Overview of Mutual Funds Bcom Project.pdf
SanaAli374401
1029-Danh muc Sach Giao Khoa khoi 6.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
QucHHunhnh
Application orientated numerical on hev.ppt
Application orientated numerical on hev.ppt
RamjanShidvankar
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptx
AreebaZafar22
Holdier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdf
agholdier
The basics of sentences session 2pptx copy.pptx
The basics of sentences session 2pptx copy.pptx
heathfieldcps1
APM Welcome, APM North West Network Conference, Synergies Across Sectors
APM Welcome, APM North West Network Conference, Synergies Across Sectors
Association for Project Management
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
RAM LAL ANAND COLLEGE, DELHI UNIVERSITY.
Kürzlich hochgeladen
(20)
Seal of Good Local Governance (SGLG) 2024Final.pptx
Seal of Good Local Governance (SGLG) 2024Final.pptx
Z Score,T Score, Percential Rank and Box Plot Graph
Z Score,T Score, Percential Rank and Box Plot Graph
psychiatric nursing HISTORY COLLECTION .docx
psychiatric nursing HISTORY COLLECTION .docx
Ecological Succession. ( ECOSYSTEM, B. Pharmacy, 1st Year, Sem-II, Environmen...
Ecological Succession. ( ECOSYSTEM, B. Pharmacy, 1st Year, Sem-II, Environmen...
Gardella_Mateo_IntellectualProperty.pdf.
Gardella_Mateo_IntellectualProperty.pdf.
Unit-V; Pricing (Pharma Marketing Management).pptx
Unit-V; Pricing (Pharma Marketing Management).pptx
Nutritional Needs Presentation - HLTH 104
Nutritional Needs Presentation - HLTH 104
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdf
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
fourth grading exam for kindergarten in writing
fourth grading exam for kindergarten in writing
Measures of Dispersion and Variability: Range, QD, AD and SD
Measures of Dispersion and Variability: Range, QD, AD and SD
SOCIAL AND HISTORICAL CONTEXT - LFTVD.pptx
SOCIAL AND HISTORICAL CONTEXT - LFTVD.pptx
An Overview of Mutual Funds Bcom Project.pdf
An Overview of Mutual Funds Bcom Project.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
Application orientated numerical on hev.ppt
Application orientated numerical on hev.ppt
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptx
Holdier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdf
The basics of sentences session 2pptx copy.pptx
The basics of sentences session 2pptx copy.pptx
APM Welcome, APM North West Network Conference, Synergies Across Sectors
APM Welcome, APM North West Network Conference, Synergies Across Sectors
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
Social network analysis part ii
1.
Social Network Analysis Fundamental
Concepts in Social Network Analysis (Part 2) Katarina Stanoevska-Slabeva, Miriam Meckel, Thomas Plotkowiak
2.
Agenda 1. Intro 2. Measuring
Networks – Embedding Measures (Ties) – Positions and Roles (Nodes) – Group Concepts 3. Network Mechanisms 4. Network Theories © Thomas Plotkowiak 2010
3.
Introduction Knoke information exchange
network In 1978, Knoke & Wood collected data from workers at 95 organizations in Indianapolis. Respondents indicated with which other organizations their own organization had any of 13 different types of relationships. The exchange of information among ten organizations that were involved in the local political economy of social welfare services in a Midwestern city. © Thomas Plotkowiak 2010
4.
2. Network Measures 2.1
Network Measures for Actors Embedding Measures
5.
Embedding Measures •
Reciprocity (Dyad Census) • Transitivity (Triad Census) • Clustering • Density • Group-external and group-internal Ties • Other Network Mechanisms © Thomas Plotkowiak 2010
6.
Reciprocity •
With symmetric data two actors are either connected or not. • With directed data there are four possible dyadic relationships: – A and B are not connected – A sends to B – B sends to A – A and B send to each other. © Thomas Plotkowiak 2010
7.
Reciprocity II •
What is the reciprocity in this network? – Answer 1: % of pairs that have reciprocated ties / all possible pairs • AB of {AB,AC,BC} = 0.33 – Answer2: % of pairs that have reciprocated ties / existing pairs • AB of {AB,BC} = 0.5 – Answer 3: % directed ties / all directed ties • {AB,BA} of {AB, BA, AC, CA, BC, CA} = 0.33 © Thomas Plotkowiak 2010
8.
Transitivity •
With undirected there are four possible types of triadic relations – No ties – One tie – Two Ties – Three Ties • The count of the relative prevalence of these four types of relations is called "triad census“. A population can be characterized by: – "isolation" – "couples only" – "structural holes" (one actors is connected to two others, who are not connected to each other) – or "clusters" © Thomas Plotkowiak 2010
9.
Transitivity II
Directed Networks M-A-N number: M # of mutual positive dyads A #asymmetric dyads N #of null dyads D =Down, U = Up, C = Cyclic, T= Transitive © Thomas Plotkowiak 2010
10.
Triad Census Models
(all) (all) Linear Hierarchy Model Every triad is 030T (all) (all) (all) Balance Model with Two Cliques (Heider Balance) Triads either 300 or 102 Ranked Clusters Model (Hierarchy of Cliques) Triads: 300, 102, 003, 120D, 120U, 030T, 021D, 021U © Thomas Plotkowiak 2010
11.
Example
Directed information exchange network 9 8 7 6 1 5 3 10 2 4 The exchange of information among ten organizations that were involved in the local political economy of social welfare services in a Midwestern city. © Thomas Plotkowiak 2010
12.
A Transitivity III
1 3 B C 2 • How to measure transitivity? – A) Divide the number of found transitive triads by the total number of possible triplets (for 3 nodes there are 6 possibilities) – B) Norm the number of transitive triads by the number of cases where a single link could complete the triad. Norm {AB, BC, AC} by {AB, BC, anything) (for 3 nodes there are 4 possibilities) © Thomas Plotkowiak 2010
13.
Transitivity IV
146/720 146/217 © Thomas Plotkowiak 2010
14.
Clustering Most actors live
in local neighborhoods and are connected to one another. A large proportion of the total number of ties is highly "clustered" into local neighborhoods. VS. © Thomas Plotkowiak 2010
15.
Global clustering coefficient
Closed triplet Triplet © Thomas Plotkowiak 2010
16.
Average Local Clustering
coefficient A measure to calculate how clustered the graph is we examine the local neighborhood of an actor (all actors who are directly connected to ego) and calculate the density in this neighborhood (leaving out the ego). After doing this for all actors, we can characterize the degree of clustering as an average of all the neighborhoods. C=1 C = 1/3 C=0 © Thomas Plotkowiak 2010
17.
Individual local clustering
coefficient (in this case for directed ties) Clustering can also be examined for each actor: – Notice actor 6 has three neighbors and hence only 3 possible ties. Of these only one is present, so actor 6 is not highly clustered. – Actor eight has 6 neighbors and hence 15 pairs of neighbors and is highly clustered. 2 edges out of 6 edges © Thomas Plotkowiak 2010
18.
Density for groups
Instead of calculating the density of the whole network (last lecture), we can calculate the density of partitions of the network. Governmental agencies Non-governmental generalist Welfare specialists A social structure in which individuals were highly clustered would display a pattern of high densities on the diagonal, and low densities elsewhere. © Thomas Plotkowiak 2010
19.
Density for groups
II • Group 1 has dense in and out ties to one another and to the other populations • Group 2 have out-ties among themselves and with group 1 and have high densities of in-ties with all three sub populations The density in the 1,1 block is .6667.That is, of the six possible directed ties among actors 1, 3, and 5, four are actually present The extend of how those blocks characterize all the individuals within those blocks can be assessed by looking at the standard deviations. The standard deviations measure the lack of homogeneity within the partition, or the extent to which the actors vary. © Thomas Plotkowiak 2010
20.
E-I Index •
The E-I (external – internal) index takes the number of ties of group members to outsiders, subtracts the number of ties to other group members, and divides by the total number of ties. (1-4)/7 = -3/7 (1-2)/7 = -1/7 © Thomas Plotkowiak 2010
21.
E-I Index II •
The resulting E-I index ranges from -1 (all ties internal) to +1 (all ties external). Ties between members of the same group are ignored. • The E-I index can be applied at three levels: – entire population – each group – each individual Notice: The relative size of sub populations (e.g. 10 vs. 1000) have dramatic consequences for the degree of internal and external contacts, even when individuals may choose contacts at random. © Thomas Plotkowiak 2010
22.
E-I Index for
groups Notice that the data has been symmetrized © Thomas Plotkowiak 2010
23.
E-I Index for
the entire population Notice that the data has been symmetrized Internal: 7*2/64 = 21% External 25*2/64 = 70% E-I (50-14)/64 = 56% © Thomas Plotkowiak 2010
24.
Permutation Tests To assess
whether the E-I index value is significantly different that what would be expected by random mixing a permutation test is performed. Notice: Under random distribution, the E-I Index would be expected to have a value of .467 which is not much different from .563, especially given the standard error .078 (given the result the difference of .10 could be just by chance) © Thomas Plotkowiak 2010
25.
E-I Index for
individuals Notice: Several actors (4,6,9) tend toward closure , while others (10,1) tend toward creating ties outside their groups. © Thomas Plotkowiak 2010
26.
2. Network Measures 2.2
Network Measures for Actors Position & Roles
27.
Positions & Roles •
Structural Equivalence • Automorphic Equivalence • Regular Equivalence • Measuring similarity/dissimilarity • Visualizing similarity and distance • Measuring automorphic equivalence • Measuring regular equivalence • Blockmodelling © Thomas Plotkowiak 2010
28.
Chinese Kinship Relations
© Thomas Plotkowiak 2010
29.
Positions and Roles •
Positions: Actors that show a similar structure of relationships and are thus similarly embedded into the network. • Roles: The pattern of relationships of members of same or different positions. • Note: Many of the category systems used by sociologists are based on "attributes" of individual actors that are common across actors. © Thomas Plotkowiak 2010
30.
Similarity •
The idea of "similarity" has to be rather precisely defined • Nodes are similar if they fall in the same "equivalence class" – We could come up with a equivalence class of out-degree of zero for example • There are three particular definitions of equivalence: – Strucutral Equivalence – Automorphic Equivalence (rarely used) – Regular Equivalence © Thomas Plotkowiak 2010
31.
Strucutral Equivalence •
Structural Equivalence: Two structural equivalent actors could exchange their positions in a network without changing their connections to the other actors in the network. • Structural equivalence is the "strongest" form of equivalence. • Problem: Imagine two teachers in Toronto and St. Gallen. Rather than looking for connections to exactly the same persons we would like to find connection to similar persons but not exactly the same ones. © Thomas Plotkowiak 2010
32.
Automorphic Equivalence •
Automorphic Equivalence: Two persons could change their positions in the network, without changing the structure of the network (Notice that after the exchange they would be partially connected to other persons than before) • Problem: How big do we have to define the radius in which we analyze the structure of the network (1, 2, 3 … steps) • For the One-Step Radius we consider the NUMBER of: – asymetric outgoing, – asymetric incoming, – symetric in- and outgoing, – and not existing ties. © Thomas Plotkowiak 2010
33.
1 Step, 2
Step Equivalence ? 1 2 © Thomas Plotkowiak 2010
34.
Regular Equivalence •
Regular Equivalence: Two positions are considered as similar, if every important Aspect of the observed structure applies (or does not apply)for both positions. • For the One-Step Radius we consider the EXISTENCE of : – asymetric outgoing, – asymetric incoming, – symetric in- and outgoing, – and not existing ties. © Thomas Plotkowiak 2010
35.
1
A B and C are B C regular equivalent D E F G H 2 A 3 A B and C are B and C are B C B C automorph structural equivalent equivalent D E F G H D E F G H © Thomas Plotkowiak 2010
36.
Computing Positional Similarity Example
Information exchange network © Thomas Plotkowiak 2010
37.
Measuring Similarity
Adjacency Matrix 1 Coun 2 Comm 3 Educ 4 Indu 5 Mayr 6 WRO 7 News 8 UWay 9 Welf 10 West 1 Coun --- 1 0 0 1 0 1 0 1 0 2 Comm 1 --- 1 1 1 0 1 1 1 0 3 Educ 0 1 --- 1 1 1 1 0 0 1 4 Indu 1 1 0 --- 1 0 1 0 0 0 5 Mayr 1 1 1 1 --- 0 1 1 1 1 6 WRO 0 0 1 0 0 --- 1 0 1 0 7 News 0 1 0 1 1 0 --- 0 0 0 8 UWay 1 1 0 1 1 0 1 --- 1 0 9 Welf 0 1 0 0 1 0 1 0 --- 0 10 West 1 1 1 0 1 0 1 0 0 --- © Thomas Plotkowiak 2010
38.
Measuring Similarity Concatenated Row
& Colum View 1 Coun 2 Comm 3 Educ 4 Indu 5 Mayr 6 WRO 7 News 8 UWay 9 Welf 10 West --- 1 0 1 1 0 0 1 0 1 1 --- 1 1 1 0 1 1 1 1 0 1 --- 0 1 1 0 0 0 1 0 1 1 --- 1 0 1 1 0 0 1 1 1 1 --- 0 1 1 1 1 0 0 1 0 0 --- 0 0 0 0 1 1 1 1 1 1 --- 1 1 1 0 1 0 0 1 0 0 --- 0 0 1 1 0 0 1 1 0 1 --- 0 0 0 1 0 1 0 0 0 0 --- --- 1 0 0 1 0 1 0 1 0 1 --- 1 1 1 0 1 1 1 0 0 1 --- 1 1 1 1 0 0 1 1 1 0 --- 1 0 1 0 0 0 1 1 1 1 --- 0 1 1 1 1 0 0 1 0 0 --- 1 0 1 0 0 1 0 1 1 0 --- 0 0 0 1 1 0 1 1 0 1 --- 1 0 0 1 0 0 1 0 1 0 --- 0 1 1 1 0 1 0 1 0 0 --- © Thomas Plotkowiak 2010
39.
Pearson correlation coefficients,
covariances and cross-products • Person correlation (ranges from -1 to +1) summarize pair- wise structural equivalence. © Thomas Plotkowiak 2010
40.
Pairwise Structural Equivalence
We can see, for example, that 9 node 1 and node 9 have identical patterns of ties. 8 The Pearson correlation measure does not pay attention to the overall 7 prevalence of ties (the mean 6 of the row or column), and it 1 does not pay attention to differences between actors in 5 3 the variances of their ties. Often this is desirable to 10 focus only on the pattern, 2 4 rather than the mean and variance as aspects of similarity between actors. © Thomas Plotkowiak 2010
41.
Euclidean squared distances Euclidean
or squared Euclidean distances are not sensitive to the linearity of association and can be used with valued or binary data. Other similar measures can be Jaccard or hamming distance. © Thomas Plotkowiak 2010
42.
Going from pairs
to groups of structural equivalence It is often useful to examine the similarities or distances to try to locate groupings of actors (that is, larger than a pair) who are similar. By studying the bigger patterns of which groups of actors are similar to which others, we may also gain some insight into "what about" the actor's positions is most critical in making them more similar or more distant. In the next two sections we will cover how multi-dimensional scaling and hierarchical cluster analysis can be used to identify patterns in actor-by-actor similarity/distance matrices. Both of these tools are widely used in non-network analysis; there are large and excellent literatures on the many important complexities of using these methods. Our goal here is just to provide just a very basic introduction. © Thomas Plotkowiak 2010
43.
Hierarchical Clustering •
Hierarchical Clustering: – Initially places each case in its own cluster – The two most similar cases are then combined – This process is repeated until all cases are agglomerated into a single cluster (once a case has been joined it is never re-classsified) © Thomas Plotkowiak 2010
44.
Multi Dimensional Scaling •
MDS represents the patterns of similarity or dissimilarity in the profiles among the actors as a "map" in a multi- dimensional space. This map lets us see how "close" actors are and whether they "cluster". – Stress is a measure of badness of fit – The author has to determine the meaning of the dimensions © Thomas Plotkowiak 2010
45.
Finding automorphic equivalence (for
binary data) • Brute Force Approach: All the nodes of a graph are exchanged and the distances among all pairs of actors in the new graph are compared to the original one. When the new and the old graph have the same distances among nodes the "swapping" that was done identified the automorphic position. • Brute Force is expensive (363880 Permutations!!) © Thomas Plotkowiak 2010
46.
Regular Equivalence Block Matrix Informal
Definition: Two actors are regularly equivalent if they have similar patterns of ties to equivalent others. Problem: Each definition of each position depends on its relations with other positions. Where to start? Sender Repeater Receiver © Thomas Plotkowiak 2010
47.
Regular Equivalence Block Matrix
Block Image • Create a matrix so that each actor in each partition has the same pattern of connection to actors in the other partition. – Notice: We don’t care about ties among members of the same regular class! – A sends to {BCD} but none of {EFGHI} – {BCD} does not send to A but to {EFGHI} – {EFGHI} does not send to A or {BCD} A B C D E F G H I A --- 1 1 1 0 0 0 0 0 B 0 --- 0 0 1 1 0 0 0 C 0 0 --- 0 0 0 1 0 0 A B,C,D E,F,G,H,I D 0 0 0 --- 0 0 0 1 1 A --- 1 0 E 0 0 0 0 --- 0 0 0 0 F 0 0 0 0 0 --- 0 0 0 B,C,D 0 --- 1 G 0 0 0 0 0 0 --- 0 0 E,F,G,H,I 0 0 --- H 0 0 0 0 0 0 0 --- 0 I 0 0 0 0 0 0 0 0 --- © Thomas Plotkowiak 2010
48.
Algorithms for detection
of Regular Equivalence Tabu Search • This method of blocking and relies on extensive use of the computer. Tabu search is trying to implement the same idea of grouping together actors who are most similar into a block. • Tabu search does this by searching for sets of actors who, if placed into a blocks, produce the smallest sum of within-block variances in the tie profiles. • If actors in a block have similar ties, their variance around the block mean profile will be small. • So, the partitioning that minimizes the sum of within block variances is minimizing the overall variance in tie profiles © Thomas Plotkowiak 2010
49.
Algorithms for detection
of Regular Equivalence Tabu Search Results 9 (2,5) for example, are pure "repeaters" 8 7 6 1 5 3 10 2 4 The set { 6, 10, 3 } send to only two other types (not all three other types) and receive from only one other type. © Thomas Plotkowiak 2010
50.
Blockmodeling Blockmodeling is able
to include all kinds of equivalences into one analysis Examples of blocks: • Complete blocks (everybody is connected with each other inside the block) • Null blocks (people in this block are not connected to anybody) • Regular blocks, people share the same regular equivalence class in this block © Thomas Plotkowiak 2010
51.
Blockmodels Matrix Permutation
© Thomas Plotkowiak 2010
52.
Blockmodels Student Government. Discussion
relation among the eleven students who were members of the student government at the University of Ljubljana in Sloveninia. The students were asked to indicate with whom of their fellows they discussed matters concerning the administration of the university informally. © Thomas Plotkowiak 2010
53.
General Blockmodelling with
predefined partitions © Thomas Plotkowiak 2010
54.
Blockmodeling based on
actors-attributes © Thomas Plotkowiak 2010
55.
Blockmodels Matrix Representation
© Thomas Plotkowiak 2010
56.
Blockmodels Matrix Permutation
© Thomas Plotkowiak 2010
57.
2. Network Measures 2.2
Network Measures Subgroups Cohesive Subgroups
58.
Cohesive Subgroups Cohesive subgroups:
We hypothesize that cohesive subgroups are the basis for solidarity, shared norms, identity and collective behavior. Perceived similarity, for instance, membership of a social group, is expected to promote interaction. We expect similar people to interact a lot, at least more often than with dissimilar people. © Thomas Plotkowiak 2010
59.
Example – Families
in Haciendas (1948) Each arc represents "frequent visits" from one family to another. © Thomas Plotkowiak 2010
60.
Components
A semiwalk from vertex u to vertex v is a sequence of lines such that the end vertex of one line is the starting vertex of the next line and the sequence starts at vertex u and end at vertex v. A walk is a semiwalk with the additional condition that none of its lines are an arc of which the end vertex is the arc's tail Note that v5 v3 v4 v5 v3 is also a walk to v3 © Thomas Plotkowiak 2010
61.
Paths A semipath is
a semiwalk in which no vertex in between the first and last vertex of the semiwalk occurs more than once. A path is a walk in which no vertex in between the first and last vertex of the walk occurs more than once. © Thomas Plotkowiak 2010
62.
Connectedness
A network is (weakly) connected if each pair of vertices is connected by a semipath. A network is strongly connected if each pair of vertices is connected by a path. This network is not connected because v2 is isolated. © Thomas Plotkowiak 2010
63.
Connected Components
A (weak) component is a maximal (weakly) connected subnetwork. A strong component is a maximal strongly connected subnetwork. v1,v3,v4,v5 are a weak component v3,v4,v5 are a strong component © Thomas Plotkowiak 2010
64.
Example Strong Components 1.
Net > Components > {Strong, Weak} © Thomas Plotkowiak 2010
65.
Cliques and Complete
Subnetworks A clique is a maximal complete subnetwork containing three vertices or more. (cliques can overlap) v2,v4,v5 is not a clique v1,v6,v5 is a clique v2,v3,v4,v5 is a clique © Thomas Plotkowiak 2010
66.
n-Clique & n-Clan n-Clique:
Is a maximal complete subgraph, in the analyzed graph, each node has maximally the distance n. A Clique is a n-Clique with n=1. n-Clan: Ist a maximal complete subgraph, where each node has maximally the distance n in the resulting graph 2-Clique 2-Clan © Thomas Plotkowiak 2010
67.
n-Clans & n-Cliques
6 5 1 4 2 3 2-Clans: 123,234,345,456,561,612 2-Cliques: 123,234,345,456,561,612 and 135,246 © Thomas Plotkowiak 2010
68.
k-Plexes k-Plex: A k-Plex
is a maximal complete subgraph with gs nodes, in which each node has at least connections with gs-k nodes. 6 5 1 4 2 3 2-Plexe:s 1234, 2345, 3456, 4561, 5612, 6123 In general k-Plexes are more robust than Cliques und Clans. © Thomas Plotkowiak 2010
69.
Overview Subgroups
4 3 4 3 4 3 1 2 1 2 1 2 2 Components 1 Component 1 Component 2 2-Clans (341,412) 1 2-Clans (124) 2 2-Cliques (341,412) 1 2-Clique (124) 4 3 4 3 1 Component 1 Component 1 2-Clan (1234) 1 2-Clan (1234) 1 2-Clique (1234) 1 2-Clique (1234) 1 2-Plex (1234) 1 2-Plex (1234) 1 2 1 2 1 Clique © Thomas Plotkowiak 2010
70.
Overview Groupconcepts •
1-Clique, 1-Clan und 1-Plex are identical • A n-Clan is always included in a higher order n-Clique Component 2-Clique 2-Clan 2-Plex Clique © Thomas Plotkowiak 2010
71.
k-Cores A •k-core is
a maximal subnetwork in which each vertex has at Net > Components > {Strong, Weak} least degree k within the subnetwork. © Thomas Plotkowiak 2010
72.
k-Cores k-cores are nested
which means that a vertex in a 3-core is also part of a 2-core but not all members of a 2-core belong to a 3- core. © Thomas Plotkowiak 2010
73.
k-Cores Application •
K-cores help to detect cohesive subgroups by removing the lowes k-cores from the network until the network breaks up into relatively dense components. • Net > Partitions > Core >{Input, Output, All} © Thomas Plotkowiak 2010
74.
© Thomas Plotkowiak
2010
75.
3. Network Mechanisms
76.
Network Mechanisms •
Tie Outdegree Effect • In/Out Popularity Effect • Reciprocity • In/Out Activity Effect • Transitivity • In/Out Assortativity Effect & Three-Cycles Effect • Covariate Similarity Effect • Balance Effect • Covariate Ego-Effect • Covariate Alter-Effect • Same Covariate Effect © Thomas Plotkowiak 2010
77.
Outdegree Effect •
The most basic effect is defined by the outdegree of actor i. It represents the basic tendency to have ties at all, • In a decision-theoretic approach this effect can be regarded as the balance of benefits and costs of an arbitrary tie. • Most networks are sparse (i.e., they have a density well below 0.5) which can be represented by saying that for a tie to an arbitrary other actor – arbitrary meaning here that the other actor has no characteristics or tie pattern making him/her especially attractive to i –, the costs will usually outweigh the benefits. Indeed, in most cases a negative parameter is obtained for the outdegree effect. © Thomas Plotkowiak 2010
78.
Reciprocity Effect •
Another quite basic effect is the tendency toward reciprocity, represented by the number of reciprocated ties of actor i. This is a basic feature of most social networks (cf. Wasserman and Faust, 1994, Chapter 13) i j © Thomas Plotkowiak 2010
79.
Transitivity and other
triadic effects • Next to reciprocity, an essential feature in most social networks is the tendency toward transitivity, or transitive closure (sometimes called clustering): friends of friends become friends, or in graph-theoretic terminology: two-paths tend to be, or to become, closed (e.g., Davis 1970, Holland and Leinhardt 1971). j j i i h h Transitive triplet Three cycle © Thomas Plotkowiak 2010
80.
Balance Effect •
An effect closely related to transitivity is balance (Newcomb, 1962), which is the same as structural equivalence with respect to out-ties (Burt, 1982), is the tendency to have and create ties to other actors who make the same choices as ego. A D B C © Thomas Plotkowiak 2010
81.
In/Out Popularity Effect •
The degree-related popularity effect is based on indegree or outdegree of an actor. Nodes with higher indegree, or higher outdegree, are more attractive for others to send a tie to. • That implies that high indegrees reinforce themselves, which will lead to a relatively high dispersion of the indegrees (a Matthew effect in popularity as measured by indegrees, cf. Merton, 1968 and Price, 1976). A B C D © Thomas Plotkowiak 2010
82.
In/Out Activity Effect •
Nodes with higher indegree, or higher outdegree respectively, will have an extra propensity to form ties to others. • The outdegree-related activity effect again is a self-reinforcing effect: when it has a positive parameter, the dispersion of outdegrees will tend to increase over time, or to be sustained if it already is high. A B C D © Thomas Plotkowiak 2010
83.
Preferential Attachment •
Notice: These four degree-related effects can be regarded as the analogues in the case of directed relations of what was called cumulative advantage by Price (1976) and preferential attachment by Barabasi and Albert (1999) in their models for dynamics of non-directed networks: a self-reinforcing process of degree differentiation. © Thomas Plotkowiak 2010
84.
In/Out Assortativity Effect •
Preferences of actors dependent on their degrees. Depending on their own out- and in-degrees, actors can have differential preferences for ties to others with also high or low out- and in-degrees (Morris and Kretzschmar 1995; Newman 2002) A D B C E F © Thomas Plotkowiak 2010
85.
Covariate Similarity Effect •
The covariate similarity effect, describes whether ties tend to occur more often between actors with similar values on a value (homophily effect). Tendencies to homophily constitute a fundamental characteristic of many social relations, see McPherson, Smith-Lovin, and Cook (2001). • Example: Ipad Owners tend to be friends with other Ipad owners. © Thomas Plotkowiak 2010
86.
Covariate Ego Effect •
The covariate ego effect, describes that actors with higher values on a covariate tend to nominate more friends and hence have a higher outdegree. • Example: Heavier smokers have more friends. © Thomas Plotkowiak 2010
87.
Covariate Alter Effect •
The alter effect describes whether actors with higher V values will tend to be nominated by more others and hence have higher indegrees. • Example: Beautiful people have more friends. © Thomas Plotkowiak 2010
88.
Modeling networks 1. Actor
Based modeling for longitudonal data – SIENA (analysis of repeated measures on social networks and MCMC- estimation of exponential random graphs) 2. Stochastic modeling for panels – Pnet objective function Model 1 Model 2 Model3 esti estim s.e. p estim s.e p s.e. p m outdegree (density) -2,46 0,12 <0,0001* -4,04 0,23 <0,0001* -1,99 0,13 <0,0001* reciprocity 2,57 0,20 <0,0001* 2,29 0,22 <0,0001* 3,02 0,21 <0,0001* transitive triplets 0,07 0,01 <0,0001* transitive mediated triplets -0,03 0,01 0,0005* transitive ties 1,47 0,24 <0,0001* 3-cycles -0,06 0,02 0,0037* attribute party 1,13 0,15 <0,0001* 0,73 0,15 <0,0001* attribute gender -0,11 0,15 0.48 © Thomas Plotkowiak 2010
89.
3. Network Theories Homophily
& Assortativity Power Laws & Preferential Attachment The Strength of Weak Ties Small Worlds Social Capital
90.
3.1Homophily
91.
Homophily •
Homophily (i.e., love of the same) is the tendency of individuals to associate and bond with similar others. (Mechanisms of selection vs influence) • In the study of networks, assortative mixing is a bias in favor of connections between network nodes with similar characteristics. In the specific case of social networks, assortative mixing is also known as homophily. The rarer disassortative mixing is a bias in favor of connections between dissimilar nodes. Low Homophily High Homophily © Thomas Plotkowiak 2010
92.
Homophily II Types (acc.
to McPherson et. Al 2001): – Race and Ethnicity (Marsden 1987, 88| Louch 2000, Kalleberg et al 1996, Laumann 1973…) – Sex and Gender (Maccoby 1998, Eder & Hallinan 1978, Shrum et al 1988, Huckfeldt & Sprague 1995, Brass 1985 …) – Age (Fischer 1977,82, Feld 1982, Blau et Al 1991, Burt 1990,91…) – Religion (Laumann 1973, Verbrugge 1977, Fischer 1977,82, Marsden 1988, Louch 2000…) – Education, Occupation and Social Class (Laumann 1973, Marsden 1987, Verbrugge 1977, Wright 1997, Kalmijn 1998…) – Network Positions (Brass 1985, Burt 1982, Friedkin 1993…) – Behavior (Cohen 1977, Kandel 1978, Knocke 1990…) – Attitudes, Abilities, Beliefs and Aspirations (Jussim & Osgood 1989, Huckfeldt & Sprague 1995, Verbrugge 1977,83, Knocke 1990) © Thomas Plotkowiak 2010
93.
Schellings Segregation Demo
© Thomas Plotkowiak 2010
94.
3.2 Power Laws
& Preferential Attachment
95.
Power Law distribution •
As a function of k, what fraction of pages on the Web have k in-links? • A natural guess the normal, or Gaussian, distribution • Central Limit Theorem (roughly): if we take any sequence of small independent random quantities, then in the limit their sum will be distributed according to the normal distribution © Thomas Plotkowiak 2010
96.
Power Law distribution But
when people measured the Web, they found something very different: The fraction of Web pages that have k in-links is approximately proportional to 1/k^2 • Power law function • Popularity exhibits extreme imbalances: there are few very popular Web pages that have extremely many in-links True for other domains: • the fraction of telephone numbers that receive k calls per day: 1/k^2 • the fraction of books bought by k people: 1/k^3 • the fraction of scientific papers that receive k citations: 1/k^3 © Thomas Plotkowiak 2010
97.
Preferential attachment leads
to power laws • A preferential attachment process is any of a class of processes in which some quantity, typically some form of wealth or credit, is distributed among a number of individuals or objects according to how much they already have, so that those who are already wealthy receive more than those who are not. Notice: "Preferential attachment" (A.L. Barabasi and R.Albert 1999) is only the most recent of many names that have been given to such processes. • Notice: Preferential attachment can, under suitable circumstances, generate power law distributions. © Thomas Plotkowiak 2010
98.
Preferential Attachment Demo DEMO
with NETLOGO © Thomas Plotkowiak 2010
99.
3.3 Balance Theory
100.
Balance Theory Franz Heider Franz
Heider (1940): A person (P) feels uncomfortable whe he ore she disagrees with his ore her friend(O) on a topic (X). P feels an urge to change this imbalance. He can adjust his opinion, change his affection for O, or convince himself that O is not really opposed to X. © Thomas Plotkowiak 2010
101.
Balance Theory (a) +
+ + : three people are mutual friends (c) - - + : two people are friends, and they have mutual enemy in the third (b) + + - : A is a friend with B and C; but B and C – enemies (d) - - - : all enemies; motivates two of them to “team up” against the third b and d represent unstable relationship © Thomas Plotkowiak 2010
102.
Balance Theory
Community in a New England Monastery Young Turks (1), Loyal Opposition (2), Outcasts (3) Interstitial Group (4) © Thomas Plotkowiak 2010
103.
Balance Theory International Relations
© Thomas Plotkowiak 2010
104.
3.4 Strength of
weak ties
105.
Strength of Weak
Ties Mark Granovetter • “One of the most influential sociology papers ever written” (Barabasi) – One of the most cited (Current Contents, 1986) • Accepted by the American Journal of Sociology after 4 years of unsuccessful attempts elsewhere. • Interviewed people and asked: “How did you find your job?” – Kept getting the same answer: “through an acquaintance, not a friend” © Thomas Plotkowiak 2010
106.
Basic Argument •
Classify interpersonal relations as “strong”, “weak”, or “absent” • Strength is (vaguely) defined as “a (probably linear) combination of… – the amount of time, – the emotional intensity, – the intimacy (mutual confiding), – and the reciprocal services which characterize the tie • The stronger the tie between two individuals, the larger the proportion of people to which they are both tied (weakly or strongly) © Thomas Plotkowiak 2010
107.
Strong Ties •
If person A has a strong tie to both B and C, then it is unlikely for B and C not to share a tie. A B C © Thomas Plotkowiak 2010
108.
Weak Ties for
Information Diffusion „Intuitively speaking, this means that whatever is to be diffused can reach a larger number of people, and traverse greater social distance, when passed through weak ties rather than strong.“ © Thomas Plotkowiak 2010
109.
3.4 Small World
Phenomenon
110.
Connectivity and the
Small World 1. Travers and Milgram’s work on the small world is responsible for the standard belief that “everyone is connected by a chain of about 6 steps.” 2. Two questions: – Given what we know about networks, what is the longest path (defined by handshakes) that separates any two people? – Is 6 steps a long distance or a short distance? © Thomas Plotkowiak 2010
111.
Example: Two Hermits
on opposite sites of the country OH Store Hermit Owner Truck Manager Driver Corporate Corporate Manager President Congress Congress Rep. Rep. Corporate Corporate President Manager Truck Manager Driver Store Mt. Owner Hermit © Thomas Plotkowiak 2010
112.
Milgrams Test Milgram’s test:
Send a packet from sets of randomly selected people to a stockbroker in Boston. Experimental Setup: Arbitrarily select people from 3 pools: – People in Boston – Random in Nebraska – Stockholders in Nebraska © Thomas Plotkowiak 2010
113.
Results •
Most chains found their way through a small number of intermediaries. • What do these two findings tell us of the global structure of social relations? © Thomas Plotkowiak 2010
114.
Results II 1. Social
networks contains a lot of short paths 2. People acting without any sort of global ‘map’ are effective at collectively finding these short paths © Thomas Plotkowiak 2010
115.
The Watts-Strogatz model •
Two main principles explaining short paths: homophily and weak ties: • Homophily: every node forms a link to all other nodes that lie within a radius of r grid steps • Weak ties: each nodes forms a link to k other random nodes • Suppose, everyone lives on a two-dimensional grid (as a model of geographic proximity) © Thomas Plotkowiak 2010
116.
Watts-Strogatz
© Thomas Plotkowiak 2010
117.
The Watts-Strogatz model •
Suppose, we only allow one out of k nodes to a to have a single random friend • k * k square has k random links - consider it as a single node • Surprising small amount of randomness is enough to make the world “small” with short paths between every pair of nodes © Thomas Plotkowiak 2010
118.
Decentralized Search •
People are able to collectively find short paths to the designated target while they don’t know the global ‘map’ of all connections • Breadth-first search vs. tunneling • Modeling: – Can we construct a network where decentralized search succeeds? – If yes, what are the qualitative properties of such a network? © Thomas Plotkowiak 2010
119.
A model for
decentralized search • A starting node s is given a message that it must forward to a target node t • s knows only the location of t on the grid, but s doesn’t know the edges out of any other node • Model must span all the intermediate ranges of scale as well © Thomas Plotkowiak 2010
120.
Modeling the process
of decentralized search • We adapt the model by introducing clustering exponent q • For two nodes v and w, d(v,w) - the number of steps between them • Random edges now generated with probability proportional to d(v,w)-q • Model changes with different values q: – q=0 : links are chosen uniformly at random – when q is very small : long-range links are “too random” – when q is large: long-range links are “not random enough” © Thomas Plotkowiak 2010
121.
Varying clustering exponent
© Thomas Plotkowiak 2010
122.
Decentralized Search when
q=2 Experiments show that decentralized search is more efficient when q=2 (random links follow inverse-square distribution) © Thomas Plotkowiak 2010
123.
What’s special about
q=2 • Since area in the plane grows like the square of the radius, the total number of nodes in this group is proportional to d2 • the probability that a random edge links into some node in this ring is approximately independent of the value of d. • long-range weak ties are being formed in a way that’s spread roughly uniformly over all different scales of resolution Think of the postal system: country, state, city, street, and finally the street number © Thomas Plotkowiak 2010
124.
Small-World Phenomenon Conclusions I 1.
Start from a Milgram’s experiment: (1) seems there are short paths and (2) people know how to find them effectively 2. Build mathematical models for (1) and (2) 3. Make a prediction based on the models: clustering exponent q=2 4. Validate this prediction using real data from large social networks (LiveJournal, Facebook) Why do social networks arrange themselves in a pattern of friendships across distance that is close to optimal for forwarding messages to far-off targets? © Thomas Plotkowiak 2010
125.
Small-World Phenomenon Conclusions II •
If there are dynamic forces or selective pressures driving the network toward this shape, they must be more implicit, and it remains a fascinating open problem to determine whether such forces exist and how they might operate. • Robustness, Search, Spread of disease, opinion formation, spread of computer viruses, gossip,… • For example: Diseases move more slowly in highly clustered graphs • The dynamics are very non-linear -- with no clear pattern based on local connectivity. Implication: small local changes (shortcuts) can have dramatic global outcomes (think of disease diffusion) © Thomas Plotkowiak 2010
126.
Small World Construction •
Network changes from structured to random • Given 6 Billion Nodes L starts at 3 million, decreases to 4 (!) • Clustering: starts at 0.75, decreases to zero • Most important is what happens ALONG the way. © Thomas Plotkowiak 2010
127.
Small worlds demo
© Thomas Plotkowiak 2010
128.
Interactive Summary
The biggest advantage I can gain by using SNA is… The most important fact about SNA for me is… The concept that made the most sense for me today was… The biggest danger in using SNA is … If I will use SNA in the future, I will try to make sure that… If I use SNA in my next project I will use it for … I should change my perspective on networks in considering … I have changed my opinion about SNA , finding out that… I missed today that … Before attending that seminar I didn't know that … I wish we could have covered… If I forget mostly everything that learned today, I will still remember … The most important thing today for me was … © Thomas Plotkowiak 2010
129.
Thanks for your
attention! Questions & Discussion
Jetzt herunterladen