The most valuable and innovative knowledge is hard to find,
and it lies within distributed communities and networks. Locating the
right community or person who can provide us with exactly the knowledge
that we need and who can help us solve exactly the problems that
we come upon, can be an ecient way to learn forward. In this paper, we
present the details of NetLearn; a service that acts as a knowledge lter
for learning. The primary aim of NetLearn is to leverage social network
analysis and visualization techniques to help learners mine communities
and locate experts that can populate their personal learning environments.
Basic Civil Engineering first year Notes- Chapter 4 Building.pptx
NetLearn: Social Network Analysis and Visualizations for Learning
1. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-1
NetLearn: Social Network Analysis
and Visualizations for Learning
Mohamed Amine Chatti, Matthias Jarke, Theresia Devi Indriasari
RWTH Aachen University, Germany
Marcus Specht
Open University Heerlen, Netherlands
ECTEL 2009
Nice, October 1, 2009
3. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-3
Personal Learning Environments
The environment in which I learn
A more natural and learner-
centric model to learning
Put the learner at the center and
give her control over the learning
experience
Convergence of lifelong, informal
and network learning within a
learner-controlled space
Lifelong LearningLifelong Learning Informal LearningInformal Learning
Self Organized LearningSelf Organized Learning Network LearningNetwork Learning
Personal Learning EnvironmentsPersonal Learning Environments
Pedagogical Perspective
4. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-4
LMS vs. PLE
LMS PLE
Content-centric Learner-centric
Management Sharing
Pre-defined selection of tools Learner needs first, tool selection second
One-size-fits-all Personal, responsive
Formal learning Support Informal and lifelong learning support
Centralized, closed, bounded Distributed, loosely coupled, open
Structured, heavyweight, rigid Freeform, lightweight, flexible
Top-down, hierarchical Bottom-up, emergent
command&control, one-way flow of knowledge Symmetric relationships
Knowledge-push Knowledge-pull
5. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-5
PLE: From knowledge-push to knowledge-pull
Get knowledge to learners
Knowledge overload
Need for filters to help learners find quality knowledge nodes
Explicit knowledge (information) vs. Tacit knowledge (people) (Nonaka&Takeuchi, 1995)
Need for community/network mining and expertise finding mechanisms
From Scarcity to Abundance
(Seely Brown, 1999)
6. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-6
Social Network Analysis &
Visualizations
Social Network Analysis (SNA) is the quantitative study of the relationships between
individuals or organizations (Wasserman&Faust, 1994)
A Graph G = (V , E)
where V = {1, 2, …., n} is a set of nodes (vertices)
E V x V is a set of edges (arcs, links, ties)⊆
Centrality measures:
degree, closeness, and betweenness
centrality
Social Networks Visualization
Node-link diagrams
Matrix-based
7. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-7
Network Characteristics:
Degree centrality
The degree centrality of a vertex v V is simply the degree of that vertex∈
Degree of a vertex: number of incoming and outgoing edges
in-degree
out-degree
The degree centrality finds the actor with the most influence over the network
(popularity of an actor, connector, hub)
CD(Fernando) = 6
8. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-8
Network Characteristics:
Closeness centrality
Closeness centrality is defined as inverse closeness, i.e., the sum of the distances
(shortest paths) to all other vertices
Closeness centrality focuses on how close an actor is to all the other actors in the
network (finds actors with the best visibility into what is happening in the network)
Cc(Fernando) = 1/15
Cc(Andre) = Cc(Jane) = 1/14
9. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-9
Network Characteristics:
Betweenness centrality
Betweenness centrality is defined as the sum of the fractions of shortest paths
between other actors that an actor sits on
Betweenness centrality finds actors that control the information flow of the network
(broker node in the network, great influence over what flows – and does not – in the
network)
CB(Fernando) = 1/3 + 1/3 = 0.66
CB(Carol) = 2 x (1/1 + 1/1 + 1/1 + 1/1 + 2/2 + 1/1 + 1/1) = 14
10. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-10
NetLearn
NetLearn: Social Network Analysis and Visualizations for Learning
Applying social network analysis and visualizations methods for community mining
and expertise finding
Case Study:
Co-authorship Network
1000+ TEL researchers
New bibliography entries via a Plone-based interface
Keywords either manually entered or automatically generated using the ALOA
Framework (Chatti et al., 2008)
12. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-12
NetLearn Implementation
Author Mining Module
Visualization of the global co-authorship network
Node: author; Edge: co-authorship
Interactive browsing, Different layouts
Edge betweenness clustering
Computation of centrality statistics
Reflection of the Long Tail phenomenon
Keyword Mining Module
Mining communities around specific
keywords
A keyword community is a cluster densely
connected by the same keyword
Node: author; Edge: shared keyword
Community Mining Modules
13. Lehrstuhl Informatik 5
(Informationssysteme)
Prof. Dr. M. Jarke
I5-MAC-13
NetLearn Implementation
Local Author Module
Ego-centric network of an author
Keyword Community Module
Expertise finding based on keywords
Locating researchers working on a specific
topic or topics closely related to that topic
Graph – chart – community – table views
Interest Community Module
Expertise finding based on query occurrence
in title, abstract, keyword
Referral Chain Module
Chain between two researchers
Shortest path algorithm
Expertise Finding Modules