Semantic Conflict Resolution Ontology SCROL An Ontology for Detecting and Resolving Data and Schemal

1
Semantic Conflict Resolution Ontology(SCROL)An
Ontology for Detectingand Resolving Data and
Schema-levelSemantic Conflicts

• Author Sudha Ram, Member, IEEE and
  Jinsoo Park, Member, IEEE
• Presenter Gaurav Batra

2
INTRODUCTION
3
STRUCTURE OF SCROL

• The structure of SCROL is a tree.
• Also defined horizontal relations i.e.,
• sibling relations and domain value mapping
  relations

4
(No Transcript)
5
Definition1

• A Semantic Conflict Resolution Ontology(SCROL) is
  a tuple (OC, OI, RS, RM, u),
• where
• OC, OI, RS, RM and u are defined in the next few
  slides.

6
Definition 2

• OC is a distinct set of Concepts
• The oval shapes in SCROL Structure represents
  Concepts
• Leaf Concept

7
Properties of Concepts

• Name
• Definition
• Subconcept
• Subconcept-of
• Instance
• Referenced-by
• (expressed as an ordered tuple )
• where
• s is a referring schema
• o is a referring object id (oid)
• t is a object type
• l is a level of reference

8
Definition 3

• OI in tuple is a distinct set of Instances
• The rectangle shapes in SCROL Structure
  represents Instances

9
Properties of Instances

• Name
• Definition
• Instance-of
• Referenced-by
• (expressed as an ordered tuple )
• where
• s is a referring schema
• o is a referring object id (oid)
• t is a object type
• l is a level of reference

10
Definition 4

• RS refers to a sibling relationship
• RS is a relation on OC
• The dotted lines between concepts in SCROL
  Structure represents sibling relationships(RS)
• Consists of 4 types of relationships
• Disjoint relationship
• Peer relationship
• part-of relationship
• is-a relationship

11
(No Transcript)
12
Disjoint Relationship

• When two concepts are not semantically
• equivalent

13
(No Transcript)
14
Peer Relationship

• When two concepts are semantically
• equivalent
• Data Values can be transformed into
  
  
  
  
  each
  other

15
(No Transcript)
16
Part-of Relationship

• Similar to aggregation in semantic
  data models and object-oriented data
  models

17
(No Transcript)
18
Is-a Relationship

• Similar to generalization/specialization in
  semantic data models and object-oriented data
  models

19
(No Transcript)
20
Definition 5

• RM refers to a domain value mapping relationship
  or (briefly, a mapping relationship)
• RM is a relation on OI
• The dotted lines between instances in SCROL
  Structure represents mapping relationships(RM)

21
(No Transcript)
22
Purpose of RM (Mapping Relationship)

• To determine whether data values can be
  transformed into each other

23
Properties of RM (Mapping Relationship)

• Synonyms
• Homonyms

24
Definition 6

• u is the root of tuple
• It has no parent
• Exactly one u is present in tuple

25
A Methodology for Semantic Conflict Detection and
Resolution

• CREAM (Conflict Resolution Environment for
  Autonomous Mediation)
• Components of CREAM
• Schema Designer
• Schema Mapper
• Ontology Designer
• Ontology Mapper
• Semantic Mediators
• Process of Instantiating SCROL using the CREAM
  System in the next few slides

26
Defining SCROL using Ontology Designer

• Ontology Designer is used for the detection of
  semantic conflict

27
Defining Conflict Controller, Original Context
and Target Context

• Used for detecting a certain semantic conflict

28
(No Transcript)
29
Defining the Ontology Relationship and Mapping
Knowledge

• Used for determining whether two terms are
  semantically related or not.

30
Definition 7 Ontology Relationship

Knowledge(ORK)

• ORK
• (Square Meter, Acre, ),
  (City,Country,part-of).

31
(No Transcript)
32
Definition 8Ontology Mapping Knowledge(OMK)

• OMK
• (Area, REGION.area),Acre,CITY_SIZE.gross_size)
  ,
• Square Meter, COUNTY_AREA.area),(region,REGION)
  ,
• (City,CITY_SIZE), (County, COUNTY_AREA).

33
Example for Detecting and Resolving Data level
conflicts using SCROL
34
(No Transcript)
35
Contributions and Comparisons with other
Approaches

• Domain independent ontology
• Other Approaches
• Context Interchange(COIN)
• A federated approach(Loosely coupled and Tightly
  coupled)
• Cupid
• Clio

36
Conclusion and Future Research

• Presented details of an ontology, SCROL
• Use of UML to represent SCROL
• Comparing the CREAM system with other systems
  such as Cupid and Clio
• Other Research Areas include
• Digital Libraries
• Electronic Commerce(EC)
• Multiagent systems(MAS)

37
Questions?