Mining Graph Data

COURSE INTRODUCTION

Teacher: Le Ngoc Thanh

Email: [email protected]

1
Contents
• Introduction of subjects and topics
• Rule
• Networks and graphs
• Graph mining
• Maths and applications

2
About the subject
• Subject name : Mining Graph Data
• Theory/Practice: 45/30.
• Rate of listening to lectures and self-studying: 40/60.
• References :
• Lecture slides
• Aggarwal, Charu C., and Haixun Wang, eds. Managing and mining graph
data. Vol. 40. New York: Springer, 2010.
• Easley, David, and Jon Kleinberg. "Networks, crowds, and markets:
Reasoning about a highly connected world." Significance 9 (2012): 43-44.
• Ketmaneechairat, Hathairat. "Graph Mining Laws, Tools and Case Studies."
Journal of Digital Information Management 12, no. 6 (2014): 446
• …

3
Subjects
Week Subject
1 Introduction to Graph Data Mining, Algorithms and
Applications
2 Review the knowledge related to graphs and data
mining
3 Patterns in static and dynamic graphs; generate
graph.
4 Indexing Graph and Ranking
5 Mining Graph Pattern
6 Graph Classification

4
Subjects
Week Subject

7 Clustering and community detection

8 Link prediction

9 Embedding graphs

10-15 Seminar topics includes: deep learning for graphs,

graph summarization, recommendation systems,
anomaly detection, large size graphs.

5
Contents
• Introduction of subjects and topics
• Rule
• Networks and graphs
• Graph mining
• Maths and applications

6
Regulations and academic assessment
• Theory: 40%
• Final Exam: 40%

• Lab: 30%
• Seminar Project: 30%
• Regulations:
• Copy, cheats → not be allowed to take the final exam
• Actions that disrupt the classroom → not be allowed to take
the final exam.

7
Contents
• Introduction of subjects and topics
• Rule
• Networks and graphs
• Graph mining
• Maths and applications

8
Networks and graphs
• Network often used to represent the natural relationship of
objects in the real world.
• Meanwhile, graphs show the relationship generated through
the automatic process.

Social Network Graphs 9

Social Network

• Vertices: People
• Edges: Friendships
Communication network

• Vertices: People
• Edges: email exchange
Business network

▪ Vertices: Companies
▪ Edges: relationships (financial, collaboration)
Biological network

▪ Vetices: Proteins ▪ Vertices: metabolites, enzymes

▪ Edges: interactions ▪ Edges: chemical reactions
Information network (WWW)

▪ Vertices: Web Pages

▪ Edges: Links
Social network

▪ Vertices: Twitter users

▪ Edges: Follows/conversations
Contents
• Introduction of subjects and topics
• Rule
• Networks and graphs
• Graph mining
• Maths and applications

16
Why is network analysis important?
• The system is connected by many components, if we
only focus on understanding a single individual, we
cannot grasp the whole system.
• There are 2 big questions :
• What are the structural properties of the network?
• What interactive process is happening in the network?

17
Research in the net
• In the field of network analysis, people focus on studying
network behaviors such as human-to-human behavior in
social networks.
• Predict behavior based on its measurable properties.

18
Modeling network with graphs
• Networks are not separate from graphs, they can be
re-modeled as graphs and take advantage of its
theoretical foundations.

19
Mining Network’s Challenges
• Normal graph:
• Large size, very very large size (massive)
• Too sparsity/ too density
• small diameter
• dynamic

• Requires efficient algorithms for storage and

computation.

20
Contents
• Introduction of subjects and topics
• Rule
• Networks and graphs
• Graph mining
• Maths and applications

21
Maths and applications
• Graph matching

22
Maths and applications

• Graph matching

Finding proteins with the same function across

different species based on their interaction networks
23
Maths and applications

• Graph matching

Using graphs in a network to identify hidden

identities in a social network

24
Maths and applications
• Semantic class

25
Maths and applications
• Graph clustering

26
Maths and applications
• Graph clustering

Grouping to reduce graph complexity

27
Maths and applications
• Graph clustering

Community detection
28
Maths and applications
• Graph classification
• Labeling the top
• Labeling the link
• Labeling the graph/subgraph

29
Maths and applications
• Graph classification

30
Maths and applications
• Frequent pattern mining in graph)

31
Maths and applications
• Frequent pattern mining in graph

Common chemical bond string pattern

32
Maths and applications
• Frequent pattern mining in graph

Withdrawal and deposit form

33
Maths and applications
• privacy-preserving in graph
• It may not be enough to remove the identifying
information, since the information can be interpolated
from known vertices.
• How to mask identifier information without breaking the
overall structure of the graph?

34
Maths and applications
• Link prediction

35
Maths and applications
• Link prediction

36
Maths and applications

37
References
• Aggarwal, Charu C., and Haixun Wang, eds.
Managing and mining graph data

38