Download presentation
Presentation is loading. Please wait.
Published byΛυσιστράτη Αξιώτης Modified over 5 years ago
1
A Tool to Support Ontology Creation based on Incremental Mini-Ontology Merging
Zonghui Lian Supported by
2
The Problem: Ontology Creation
Information collection and analysis Concept and relationship design Iterative construction
3
TANGO: Table ANalysis for generating Ontologies
Growing Ontology New Growing ontology Table Mini-Ontology
4
TANGO: Table ANalysis for generating Ontologies
OntoMerge Growing ontology New table New mini-ontology
5
Ontology Mapping and Merging
A simple case Mini-ontology Growing ontology New growing ontology
6
Ontology Mapping and Merging
A complex case Mini-ontology Growing ontology
7
Ontology Mapping and Merging
A complex case Mini-ontology Edited mini-ontology Growing ontology
8
Ontology Mapping and Merging
A complex case New growing ontology Edited Mini-ontology Growing ontology
9
Ontology Mapping and Merging
A more complex case Mini-ontology Growing ontology Issue: Functional/Nonfunction conflict
10
Ontology Mapping and Merging
A more complex case Growing ontology Mini-ontology Edited Mini-ontology ? Issue: functional/non-functional conflict Default: The merge will make the functional relationship set non-functional. Suggestion: If this is not what is wanted, make the non-functional edge functional before merging.
11
Ontology Mapping and Merging
A more complex case Merged ontology Issue: possible redundant relationship sets: - Country has Name - Country has Language Default: These relationship sets will be removed. Suggestion: If this is not what is wanted, remove/keep relationship sets as desired.
12
Ontology Mapping and Merging
A more complex case Merged ontology New growing ontology
13
Ontology Growing Process
Mini-ontology Edited Mini-ontology Ontology merging algorithms Ontology mapping algorithms New growing ontology Merged ontology Ontology cleaning algorithms
14
OntoMerge: Framework Ontology editor OntoMerge … Create … …
Cleaning algorithms OntoMerge … Growing ontology Mapping algorithms Merging algorithms Create … … Management functions OSM has a model-equivalent language called OSM-L, which is able to present an OSM ontology as a set of statements, conversely, a set of OSM-L statement can be transferred into OSM model too.
15
Contribution A tool to support ontology mapping, merging, and cleaning (MMC) Manual MMC Enable plug-in algorithms for semi-automatic and automatic MMC TANGO: ontology creation
16
The end
17
TANGO Project An information-gathering engine to assimilate and organize knowledge
18
TANGO’s working process includes
Recognize and normalize table information Construct mini-ontologies from normalized table Discover inter-ontology mapping (UI) Merge mini-ontology into a growing ontology
19
Ontology Mapping Based on the characteristics of object sets in two ontologies Simple mapping Joint mapping Union mapping
20
Ontology Mapping Based on the number of object sets in two different ontologies The 1:1 cardinality problem The 1:n (n:1) cardinality problem The n:m cardinality problem
21
Ontology Mapping and Merging
Simple Case ==
22
Ontology Mapping Union Mapping: 1:n or n:1 1:1
23
Ontology Mapping Join Mapping 1:n or n:1 1:1
24
OntoMerge Tool Ontology editor OntoMerge
OSM has a model-equivalent language called OSM-L, which is able to present an OSM ontology as a set of statements, conversely, a set of OSM-L statement can be transferred into OSM model too.
25
Concepts: Country, Population Males,
and Population Females Relationships: Country[1] has Population Males[1:*] and Country[1] has Population Females[1:*] Concepts: Country and Population total Relationships: Country[1] has Population total[1:*] Concepts: Country, Population total, Population Males and Population Females Relationships: Country[1] has Population total[1:*]; Population Males Isa Population total; and Population Females isa Population total Country(s) SimpleMap Country(t) (Population Males, Population Females) UnionMap Population total
26
Ontology Merging Ontology merging based on join mapping ==
Concepts: Person, First Name and Last Name Relationships: Person[1] has First Name[1:*] and Person[1] has Last Name[1:*] Concepts: Person and Name Relationships: Person[1] has Name[1:*] (First Name[1](s) , Last Name[1](s)) JoinMap Name(t) Person(s) SimpleMap Person(t) Concepts: Person, Name, First Name, and Last Name Relationships: Person[1] has Name[1:*], First Name[1] isSubPartOf Name[1] and Last Name[1] is SubPartOf Name[1]
27
OntoMerge: A tool to support ontology mapping and merging based on existed algorithms Provide a framework where mapping and merging algorithms can be plugged in Provide IDS (issue/ default/ suggestions) Provide users a friendly UI and allow users to fully control mapping and merging including manually map and merge ontologies
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.