Mining Graphs with Constrains on Symmetry and Diameter

25
Mining Graphs with Constrains on Symmetry and Diameter Natalia Vanetik Deutsche Telecom Laboratories at Ben- Gurion University IWGD10 workshop July 14th, 2010 Jiuzhaigou, China 1

description

Mining Graphs with Constrains on Symmetry and Diameter. Natalia Vanetik Deutsche Telecom Laboratories at Ben-Gurion University. Graph mining (1) Problem statement. Graph mining (2) Motivation. Graphs are everywhere Chemical compounds ( Cheminformatics ) - PowerPoint PPT Presentation

Transcript of Mining Graphs with Constrains on Symmetry and Diameter

Page 1: Mining Graphs with Constrains on Symmetry and Diameter

Mining Graphs with Constrains on Symmetry and Diameter

Natalia VanetikDeutsche Telecom Laboratories at

Ben-Gurion University

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China1

Page 2: Mining Graphs with Constrains on Symmetry and Diameter

Graph mining (1)Problem statement

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China2

Page 3: Mining Graphs with Constrains on Symmetry and Diameter

Graph mining (2)Motivation• Graphs are everywhere

– Chemical compounds (Cheminformatics)– Protein structures, biological pathways/networks (Bioinformactics)– Program control flow, traffic flow, and workflow analysis – XML databases, Web, and social network analysis

• Graph is a general model– Trees, lattices, sequences, and items are degenerated graphs

• Diversity of graphs– Directed vs. undirected, labeled vs. unlabeled (edges & vertices), weighted, with

angles & geometry (topological vs. 2-D/3-D)

• Complexity of algorithms: many problems are of high complexity (NP complete or even P-SPACE !)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China3

Page 4: Mining Graphs with Constrains on Symmetry and Diameter

Graphs, graphs, everywhere

Aspirin Yeast protein interaction network

from

H.

Jeon

g e

t al N

atu

re 4

11

, 4

1

(20

01

)

Internet Co-author network

Page 5: Mining Graphs with Constrains on Symmetry and Diameter

Constraints: diameter

• Diameter d(G) of a graph G is the maximum among minimal distances between pairs of its vertices.

• d(G)=1 implies that G is complete.• d(G)= implies that G is not connected.

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China5

d(G)=1 d(G)=2 d(G)=2 d(G)=

Page 6: Mining Graphs with Constrains on Symmetry and Diameter

Constraints: symmetry• Symmetries of a graph G are determines by its automorphism

group Aut(G).• Aut(G) is a permutation group.• Largest possible automorphism group for a graph of size n is

Sn, which has order n!

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China6

Aut(G)=S5 Aut(G)=S3 Aut(G)=D5 Aut(G)=S5

Page 7: Mining Graphs with Constrains on Symmetry and Diameter

Measuring symmetry and diameter (1)

• Graph diameter is computable in polynomial time.• Automorphism group of a graph is not likely to be

computable in polynomial time.– Best known algorithm: Nauty by B. McKay, outputs a set of generators

of Aut(G).

• Intuitively, graphs with smaller diameter and higher symmetry are more interesting.

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China7

d(G)=2 d(G)=3

Page 8: Mining Graphs with Constrains on Symmetry and Diameter

Measuring symmetry and diameter (2)

• Symmetry is harder to measure.• Observation: maximum symmetry of a graph is achieved

when is automorphism group is the symmetric group of order equal to the size of a graph.

• Suggestion: measure symmetry of G as

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China8

| |

( )( )

G

Aut Gs G

S

s(G)=|S5|/|S5|=1 s(G)=|S3|/|S5|= 1/20 s(G)=|D5|/|S5|= 1/12

Page 9: Mining Graphs with Constrains on Symmetry and Diameter

Tree decomposition of a graph

• Let G=(V,E) be a graph. Tree T is called a tree decomposition of G if– Nodes of T are subsets X1,…,Xn V such that X1…Xn=V

– If node vXiXj , then every node Xk of T on the path from Xi to Xj contains v as well.

– For every edge e=(v,u) there exists i so that u,v Xi.

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China9

1 2

4 3

G

1 23 4

1 2 4

2 34

T1={{1,2,3,4}, } T2={{1,2,4},{2,3,4}},}(} 1,2,4},{2,3,4{{){

Page 10: Mining Graphs with Constrains on Symmetry and Diameter

Minimal tree decomposition

• Width of a tree decomposition T is (max i |Xi|)-1.

• Minimum width among all tree decomposition is called tree width of a graph.

• Tree width equals maximum clique size minus 1.• Tree decomposition of minimum width is called minimal tree

decomposition.• Computing minimal tree decomposition is NP-hard problem

as it contains the problem of finding all maximum cliques in a graph.

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China10

Page 11: Mining Graphs with Constrains on Symmetry and Diameter

Different tree decompositions

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China11

1 2 3

5

4

67

8

1 2 8

2 46 8

2 3 4

8 7 6

4 5 6

1 2 8

8 7 6

2 6 8

2 46

2 3 4

4 5 6

Minimal tree decomposition

Non-minimal tree decomposition

Page 12: Mining Graphs with Constrains on Symmetry and Diameter

Intuition behind the proposed algorithm1. Compute the finest tree decomposition possible for every DB

transaction under given time constraints.2. Use basic pattern growing algorithm, such as FSG or gSpan to

extend instances of frequent patterns.3. Every time an instance of a frequent pattern is extended by an

edge of a nodea. Compute its diameter and symmetry estimates based on pattern’s

position within tree decomposition of a DB transaction;b. if one of the estimates is lower than user-specified symmetry or

diameter constraints, remove patterns instance from instance list,c. otherwise, keep the instance in the list.d. If the count of instances is higher than support bound, this is a

frequent pattern.

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China12

Page 13: Mining Graphs with Constrains on Symmetry and Diameter

How does it work?

• Let T be tree decomposition of DB graph transaction t.• Let Gt be an instance of a candidate pattern.• Let TG=(VG,EG)T be minimal subtree of T containing G.

Claim 1. d(G)d(TG).

Claim 2. s(G)≤(|LAut(TG)|X VG|X\EG|!e EG

|e|!)/|G|!

where LAut is automorphism group of TG viewed as

tree where each node X is labeled by |X|.

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China13

Page 14: Mining Graphs with Constrains on Symmetry and Diameter

Example (1)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China14

1 2 3

5

4

67

8

1 2 8

8 7 6

2 6 8

2 46

2 3 4

4 5 6

Pattern instance and corresponding subtree of minimal Tree decomposition

Diameter is at least 1Diameter is at least 2

Page 15: Mining Graphs with Constrains on Symmetry and Diameter

Example (2)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China15

1 2 3

5

4

67

8

1 2 8

8 7 6

2 6 8

2 46

2 3 4

4 5 6

Pattern instance and corresponding subtree of minimal Tree decomposition

Symmetry is at most 1Symmetry is at most 2*2!*1!*1!/4!=1/6

Page 16: Mining Graphs with Constrains on Symmetry and Diameter

Properties of estimates

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China16

Page 17: Mining Graphs with Constrains on Symmetry and Diameter

The algorithm

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China17

Page 18: Mining Graphs with Constrains on Symmetry and Diameter

Correctness

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China18

Page 19: Mining Graphs with Constrains on Symmetry and Diameter

Complexity concerns

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China19

Page 20: Mining Graphs with Constrains on Symmetry and Diameter

Test results (symmetry)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China20

Page 21: Mining Graphs with Constrains on Symmetry and Diameter

Test results (symmetry)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China21

Page 22: Mining Graphs with Constrains on Symmetry and Diameter

Test results (symmetry)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China22

Page 23: Mining Graphs with Constrains on Symmetry and Diameter

Test results (diameter)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China23

Page 24: Mining Graphs with Constrains on Symmetry and Diameter

Test results (diameter)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China24

Page 25: Mining Graphs with Constrains on Symmetry and Diameter

Test results (diameter)

IWGD10 workshop July 14th, 2010 Jiuzhaigou, China25