Social Network Analysis
• Tools for SNA • Page Rank Algorithm • Hierarchical Clustering • Recommendation System based on SNA (Collaborative Filtering) • How Facebook/Amazon uses SNA for recommendations? • Two Hop degree • Dynamism in Friendship Network of CSE-B • Online Social Networks and Clusters • Influential Nodes and Their Importance • Bibliography • Question - Answer Session
By Sohom Ghosh
TOPICS:
Tools for Analyzing Social Networks
• Gephi
• Pajek
• NetworkX (Python package)
• iGraph (Python and R package)
NetworkX
• iGraph is similar.
• >>> import networkx as nx
• >>> G=nx.Graph()
• >>> G.add_edge(1,2) # default edge data=1
• >>> G.add_edge(2,3,weight=0.9) # specify edge data
• >>> print(nx.dijkstra_path(G,'a','d'))
Page Rank Algorithm
• PageRank works by counting the number and quality of links to a page to determine a rough estimate of how important the website is. The underlying assumption is that more important websites are likely to receive more links from other websites
• The original PageRank algorithm was described by Lawrence Page and Sergey Brin
in several publications. It is given by
where PR(A) is the PageRank of page A, PR(Ti) is the PageRank of pages Ti which link to page A, C(Ti) is the number of outbound links on page Ti and d is a damping factor which can be set between 0 and 1.
PR(A) = (1-d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn))
Hierarchical Clustering
• Process:-
1. After calculating “distances”, “weights” between all pairs of vertices
2. Start with all n nodes treating them as disconnected
3. Add edges between pairs one by
one in order of decreasing weights
RESULT: Nested components, where one
can take a ‘slice’ at any level of the tree.
Recommendation System Based on SNA (Collaborative Filtering)
• Collaborative Filtering is a method of making automatic predictions
about the interests of a user by collecting preferences or taste
information from many users from a perspective of collaboration. The
underlying assumption of this approach is that, if a person A has the
same opinion as that of person B on an issue, then A is more likely to
follow B’s opinion on a different issue x than to follow the opinion of a
person chosen randomly.
Ex:- Samujjal and Sayantan have 83 mutual friends in common and are
room mates. If Sayantan buys a moto E smartphone, Samujjal is most likely
to buy it in future. In this case we should recommend him this phone.
Analysis
• Mutual Friends
• Friend Circle and Network
• Interactions
• Liked Pages/Interests
• Age Group
Two Hop Degree
• Two hop degree of an item represents a node’s ability to extend its
reach beyond its immediate neighbors. Its significance lies in the
realization of influence that a node exerts on its neighbors. A node
with higher amount of two hop degree distribution form the core
part of that group.
• If two nodes are not adjacent and have at least one mutual friend,
then they are at a distance of Two Hop from each other. Our friends
of friends are at a distance two hop from us.
• E.g. “BLUE EYES” :P
Dynamism in Friendship Network of CSE-B
1
2
4
3
6
5
Protim
Arijit
Sayantan
Pramit
Rajpratim
Utsab
1st Year
7
SnG
Dynamism in Friendship Network of CSE-B
1
2
4
3
6
5
Protim
Arijit
Sayantan
Pramit
Rajpratim
Utsab
2nd Year 1st Sem
7
SnG
8 RS
9
Samujjal
1
2
4
3
6 Arijit
Sayantan
Pramit
Rajpratim
Utsab
2nd Year 2nd Sem
7
SnG
8
RS
9
Samujjal
1
2
4
3
6 Arijit
Sayantan
Pramit
Rajpratim
Utsab
3rd Year 1st Sem
7
SnG
8
RS
9
Samujjal
1
2
4
3
6 Arijit
Sayantan
Pramit
Rajpratim
Utsab
3rd Year 2nd Sem (Expected)
7
SnG
8
RS
9
Samujjal
Dynamism in Friendship Network of CSE-B
Influential Nodes and Their Importance
• “More Connected”
• “Degree”, “Betweenness” , “Closeness” more
• “Teachers” in villages.
• For “CLEAN INDIA” drive, famous Bollywood actors and celebrities
are nominated as campaigner. This is because they have lots of fans
and are influential nodes in the network.
Bibliography
• http://www.slideshare.net/gcheliotis/social-network-analysis-
3273045
• www.images.google.com
• https://www.coursera.org/course/sna
• Presentation by James Moody on Social Network Analysis at
American Sociological Association, San Francisco, August 2004
• Data Mining for Social Network Analysis, IEEE ICDM 2006, Hong
Kong, Jaideep Srivastava, Nishith Pathak, Sandeep Mane,
Muhammas A. Ahamad University of Minnesota