Falcon-AO: Results for OAEI 2007

67 %
33 %
Information about Falcon-AO: Results for OAEI 2007
Technology

Published on December 22, 2008

Author: gcheng

Source: slideshare.net

Description

Presentation given at OM-2007. It describes Falcon-AO, an automatic ontology matching tool, which participates OAEI'2007.

Falcon-AO: Results for OAEI 2007 Wei Hu, Yuanyuan Zhao, Dan Li Gong Cheng , Honghan Wu, and Yuzhong Qu School of Computer Science and Engineering, Southeast University, Nanjing 210096, China

Outline Overview Features and Capabilities OAEI 2007 Results Conclusion

Overview

Features and Capabilities

OAEI 2007 Results

Conclusion

Overview of Falcon-AO Falcon-AO is an automatic ontology matching system to help establish interoperability between (Semantic) Web applications using different while related ontologies. Matching Web ontologies described in RDF(S) and OWL Implemented in Java An open source project under the Apache 2.0 license http:// iws.seu.edu.cn /projects/matching/

Falcon-AO is an automatic ontology matching system to help establish interoperability between (Semantic) Web applications using different while related ontologies.

Matching Web ontologies described in RDF(S) and OWL

Implemented in Java

An open source project under the Apache 2.0 license

http:// iws.seu.edu.cn /projects/matching/

History of Falcon-AO – v0.3 (2005) 2005 Linguistic Matching + Graph Matching LMO + GMO

History of Falcon-AO – v0.6 (2006) 2005 Linguistic Matching + Graph Matching + Partition-based V-Doc + GMO + I-Sub + PBM 2006

History of Falcon-AO – v0.7 (2007) Linguistic Matching + Graph Matching + Partition-based V-Doc + GMO + I-Sub + PBM 2006 2005 2007

What’s New in Falcon-AO v0.7 Bug repair Optimization Open source

Bug repair

Optimization

Open source

Architecture Model Pool Parsing ontologies into in-memory models using Jena Adjusting models using coordination rules

Model Pool

Parsing ontologies into in-memory models using Jena

Adjusting models using coordination rules

Architecture (cont’d.) Matcher Library V-Doc, I-Sub: two light- weighted linguistic matchers GMO: an iterative structural matcher PBM: partition-based block matching of large ontologies

Matcher Library

V-Doc, I-Sub: two light- weighted linguistic matchers

GMO: an iterative structural matcher

PBM: partition-based block matching of large ontologies

Architecture (cont’d.) Alignment Set Generating alignments using a widely-accepted RDF/XML format Evaluating generated alignments against reference alignments

Alignment Set

Generating alignments using a widely-accepted RDF/XML format

Evaluating generated alignments against reference alignments

Architecture (cont’d.) Central Controller Configuring matching strategies Executing matchers Combining similarity in terms of the measures of the linguistic and structural comparability

Central Controller

Configuring matching strategies

Executing matchers

Combining similarity in terms of the measures of the linguistic and structural comparability

Architecture (cont’d.) Repository Storing useful data during the matching process

Repository

Storing useful data during the matching process

Outline Overview Features and Capabilities OAEI 2007 Results Conclusion

Overview

Features and Capabilities

OAEI 2007 Results

Conclusion

Features and Capabilities – V-Doc V-Doc – constructing virtual documents for ontology matching A light-weighted linguistic matcher The virtual document of a domain entity in an ontology contains: its local description, such as labels or comments its neighboring information Document similarity is calculated via the TF/IDF technique.

V-Doc – constructing virtual documents for ontology matching

A light-weighted linguistic matcher

The virtual document of a domain entity in an ontology contains:

its local description, such as labels or comments

its neighboring information

Document similarity is calculated via the TF/IDF technique.

Features and Capabilities – GMO GMO – graph matching for ontologies An iterative structural matcher Using RDF bipartite graphs to represent ontologies Computing structural similarity between domain entities and between statements in ontologies by recursively propagating similarity in the bipartite graphs Ontologies + external mappings  GMO  additional mappings

GMO – graph matching for ontologies

An iterative structural matcher

Using RDF bipartite graphs to represent ontologies

Computing structural similarity between domain entities and between statements in ontologies by recursively propagating similarity in the bipartite graphs

Ontologies + external mappings  GMO  additional mappings

Features and Capabilities – PBM PBM – partition-based block matching of large-scale ontologies Based on the divide-and-conquer idea Advantages: Avoiding OUT-OF-MEMORY Decreasing the runtime

PBM – partition-based block matching of large-scale ontologies

Based on the divide-and-conquer idea

Advantages:

Avoiding OUT-OF-MEMORY

Decreasing the runtime

Features and Capabilities – Coordination Rules Heterogeneous ways in expressing ontologies Falcon-AO implements 21 coordination rules to eliminate “useless” axioms and reduce structural heterogeneity. Removing redundant and worthless statements owl:imports Complementing statements owl:inverseOf Reconstructing the RDF List structure Replacing RDF collection vocabularies (rdf:first …) with rdfs:member

Heterogeneous ways in expressing ontologies

Falcon-AO implements 21 coordination rules to eliminate “useless” axioms and reduce structural heterogeneity.

Removing redundant and worthless statements

owl:imports

Complementing statements

owl:inverseOf

Reconstructing the RDF List structure

Replacing RDF collection vocabularies (rdf:first …) with rdfs:member

Features and Capabilities – Similarity Combination Falcon-AO develops an approach to automatically tune the thresholds of different matchers based on the measures of The linguistic comparability: # mappings found by I-Sub / min (#domain entities in two ontologies) The structural comparability (confidence): # mappings found by GMO / #mappings found by I-Sub or V-Doc General scheme Higher comparability , lower cutoff Unmatchable if both lower

Falcon-AO develops an approach to automatically tune the thresholds of different matchers based on the measures of

The linguistic comparability:

# mappings found by I-Sub / min (#domain entities in two ontologies)

The structural comparability (confidence):

# mappings found by GMO / #mappings found by I-Sub or V-Doc

General scheme

Higher comparability , lower cutoff

Unmatchable if both lower

Outline Presentation of the system Features and capabilities OAEI 2007 Results Conclusion

Presentation of the system

Features and capabilities

OAEI 2007 Results

Conclusion

Towards a Generic Practical Tool 5 min for all Benchmark 40 min Library 3 min for all 91 tasks Conference 1.2 h GEMET vs. NALT Environment GEMET vs. AGROVOC 33 min 5.75 h Food 2 min for all 4,639 tasks Directory

Outline Presentation of the system Features and capabilities OAEI 2007 Results Conclusion

Presentation of the system

Features and capabilities

OAEI 2007 Results

Conclusion

Conclusion Falcon-AO (v0.7) performs well on most of matching tasks. In the near future, we are looking forward to extending Falcon-AO to: Integrating new matchers, especially instance-based approaches Discovering mappings between ontologies and RDB schemas Supporting data transformation by deriving executable mappings from schema matching

Falcon-AO (v0.7) performs well on most of matching tasks.

In the near future, we are looking forward to extending Falcon-AO to:

Integrating new matchers, especially instance-based approaches

Discovering mappings between ontologies and RDB schemas

Supporting data transformation by deriving executable mappings from schema matching

Thanks for your attention! Any comments are welcome! Welcome to Beijing 2008!

Add a comment

Related presentations

Presentación que realice en el Evento Nacional de Gobierno Abierto, realizado los ...

In this presentation we will describe our experience developing with a highly dyna...

Presentation to the LITA Forum 7th November 2014 Albuquerque, NM

Un recorrido por los cambios que nos generará el wearabletech en el futuro

Um paralelo entre as novidades & mercado em Wearable Computing e Tecnologias Assis...

Microsoft finally joins the smartwatch and fitness tracker game by introducing the...

Related pages

CiteSeerX — Falcon-AO: Results for OAEI 2007

BibTeX @MISC{Hu_falcon-ao:results, author = {Wei Hu and Yuanyuan Zhao and Dan Li and Gong Cheng and Honghan Wu and Yuzhong Qu}, title = {Falcon-AO ...
Read more

Falcon-AO: Results for OAEI 2007 - CEUR-WS.org - CEUR ...

Falcon-AO: Results for OAEI 2007 Wei Hu, Yuanyuan Zhao, Dan Li, Gong Cheng, Honghan Wu, and Yuzhong Qu School of Computer Science and Engineering ...
Read more

Fillable Online The Results of Falcon-AO in the OAEI 2006 ...

Fillable The Results of Falcon-AO in the OAEI 2006 Campaign - CEUR-WS.org
Read more

The Results of Falcon-AO in the OAEI 2006 Campaign ...

In this paper, we briefly introduce the architecture of Falcon-AO (version 0.6) and highlight two major improvements in the current version. Falcon-AO ...
Read more

Ontology mapping: a user survey - dl.acm.org

2007 Article Bibliometrics · Downloads (6 Weeks): 1 · Downloads (12 Months): 2 · Downloads (cumulative): 2 · Citation Count: 0 Tools and ...
Read more

Relevance-Based Evaluation of Alignment Approaches: the ...

of the OAEI 2007 food task, which did not take relevance into account. We determine relevance for the mappings based on hot topics related to this task, ...
Read more

In vitro study of mapping method interactions in a name ...

2007 Article Bibliometrics · Downloads (6 Weeks): 0 · Downloads (12 Months): 0 · Downloads (cumulative): 0 · Citation Count: 0 Tools and ...
Read more

OAEI-JODS - scribd.com

OAEI-JODS - Download as PDF File (.pdf), Text File (.txt) or read online. Scribd is the world's largest social reading and publishing site. Browse Browse.
Read more

Emerging consensus in-situ

2007 Article Bibliometrics · Downloads (6 Weeks): 0 · Downloads (12 Months): 0 · Downloads (cumulative): 0 · Citation Count: 0 Tools and ...
Read more

Results of the ontology alignment evaluation initiative 2010

Results of the ontology alignment evaluation initiative 2010. Full Text: PDF: Authors: Jérôme Euzenat: INRIA & LIG, Montbonnot, France: Alfio Ferrara ...
Read more