A Structural Approach to Community-level Social Influence
Analysis Ph.D. Viva Vclav Belk
Slide 2
Context and Motivation I Our earlier study suggested
communities influence each other 2 / 25
Slide 3
Context and Motivation II Network represents flow between
actors Actor-level social influence in healthcare, innovations,
marketing, etc. Actors embedded in communities No suitable model of
community-level influence high in-degree 3 / 25
Slide 4
Research Problem and Questions Problem: measurement, analysis,
and explanation of influence between various types of social
communities Questions 1.How can we model influence between
communities? 2.How do we detect communities acting as global
authorities/hubs? 1.Can we exploit the model to maximise
information diffusion? 4 / 25
Slide 5
Q1: How can we model influence between communities? 5 / 25
Slide 6
Methodology: COIN How What centrality actors communities impact
communities membership actors communities T 6 / 25 impacts depends
on
Slide 7
Impact and Its Aggregates 7 / 25 impacts depends on communities
row impact of a community on others column impact of others on a
community diagonal independence importance = total impact of a
community on others dependence = total impact of others on a
community importance/dependence heterogeneity measured by
entropy
Slide 8
Experiments 8 / 25
Slide 9
Influence Over Time Questions: Which communities influenced a
given community over time? How do we measure that by COIN?
Hypothesis Frequent impact higher than independence indicates
influence Experiments segment data by time window find impact
higher than independence of influenced community Discussion fora
data links represent replies forum as a proxy of community 9 /
25
Slide 10
Personal Issues vs Moderators Personal Issues influenced first
by Moderators Later by a specific moderating community, PI Mods
emphasised: strong impact 10 / 25
Slide 11
Q2: How do we detect communities acting as global
authorities/hubs? 11 / 25
Slide 12
12 / 25 importance importance entropy global authorities local
authorities low widespread low Global Authorities: Widespread High
Importance
Slide 13
13 / 25 importance importance entropy Moderators Moderators:
Authority of
Slide 14
14 / 25 dependence dependence entropy hubsdriven lowwidespread
low Global Hubs: Widespread High Dependence
Slide 15
15 / 25 dependence dependence entropy After Hours: Hub of After
Hours
Slide 16
16 / 25 SAP Business One: Core Core: Hub of dependence entropy
dependence COIN integrated to SAP PULSAR
Slide 17
Cross-Community Dynamics in Science Questions How can we
measure and explain influence between scientific communities? How
does the influence relate to communitys performance? How do we
adapt COIN? Data Scientists linked by citations AI communities
defined as conferences 17 / 25
Slide 18
COIN for Scientific Communities citations as a proxy of impact
and information flow Aggregate Measures importance: how much
information flows out of the community independence: how
introspective the community is 18 / 25 citation information
flow
Slide 19
Exporters and Isolated AI Communities Hypothesis importance
indicates exporters independence and importance indicates isolated
islands 19 / 25 independence importance exportersislands mainstream
loose exporters CBR COLT IJCAI
Slide 20
Q3: Can we exploit the model to maximise information diffusion?
20 / 25
Slide 21
Influence and Information Diffusion Actor-level diffusion
maximisation problem: Which actors to target? Cross-community
diffusion maximisation problem: Which communities to target? high
in-degree 21 / 25
Slide 22
Hypothesis: product of importance and entropy identifies seed
communities that induce high overall adoption Overall adoption
estimated by a diffusion model on Four targeting strategies:
1.Impact Focus (IF) COIN 2.Greedy (GR) 3.Group In-degree (GI)
4.Random (RA) Information Diffusion Experiments IF = importance
entropy 22 / 25 Selection vs Prediction
Slide 23
COIN Optimises Information Diffusion Selection Prediction
Greedy strategy overfits Impact Focus is more robust 23 / 25
Slide 24
Summary and Future Work COIN: computational model for community
influence Communities influencing a particular community Roles of
communities: authorities vs hubs Isolated communities loosing
influence Seed communities for information diffusion General (3
systems) and extensible Tensor-based extension of COIN captures
topics Future Work May be applicable to e.g. email networks Impact
Focus may be improved by discounting overlap Sentiment-informed
community influence 24 / 25
Slide 25
Contributions proposes a solution to the problem of
measurement, analysis, and explanation of influence between
communities purely structural approach extended to capture topics
empirical analysis of 3 systems common/different phenomena first
approach to novel problem of cross-community information diffusion
Dissemination 1 journal, 3 conference, and 1 workshop papers best
poster at NUIG research day 2013 complete results, software, data,
thesis, etc. at: 25 / 25 http://belak.net/doc/2014/thesis.html
Slide 26
Personal Issues and Moderators 26
Slide 27
CBR community: isolated CBR JELIA 27
Slide 28
CBR: isolated and shrinking rising impact factor driven by
self-citations decreasing size rigid member-base CBR was unable to
attract new members and decayed Cannot be revealed by introspective
analysis 28
Slide 29
Greedy Strategy 29
Slide 30
Group In-Degree 30 GI = # links from outside
Slide 31
Topical Dimensions of Influence COIN extended to capture topics
Based on tensor algebra Better interpretability and sensitivity
Consistent with purely structural COIN Example: V-TFL Admin vs
V-TFL Discussion actors communities topics 31