SSONDE: Semantic Similarity On liNked Data Entities
Download as pptx, pdf0 likes704 views
This document describes SSONDE, a framework for calculating semantic similarity between linked data entities. It provides instructions on downloading and installing SSONDE, crawling data from data.cnr.it to use as an example, configuring SSONDE, running it to calculate similarities, and visualizing the results. The document also discusses potential optimizations and extensions to SSONDE. It concludes by asking the reader to consider how SSONDE could be useful for their research and whether they would be interested in contributing.
![Configuration file 1
{ "StoreConfiguration":{
"KindOfStore":"JENATDB",
"RDFDocumentURIs":[ ],
"TDBDirectory":"data/CNRIT/TDB-0.8.9/CNRR/"
}, List of LOD Entities URI
"InstanceConfiguration":{ Java class Implementing ListOfInputInstances
"InstanceURIsClass":"application.dataCNRIt.GetResearcherIMATIplusCoauthor"
},
"OutputConfiguration":{ Similarity Matrix CSV - JSON encoding of top n-most
"KindOfOutput":"JSONOrderedResult", similar
"NumberOfOrderedResult":20",
"FilePath":"conf/dataCNRIt/ComplexContextResearchInterest/CRRIIntPub.res.json"
}, Context Encoded in a format in-house text format/
hopefully soon in JSON
"ContextConfiguration":{
"ContextFilePath":"conf/dataCNRIt/ComplexContextResearchInterest/CCRIIntPub.ctx"
}
}
Riccardo Albertoni 10](https://image.slidesharecdn.com/ssondedemotoshare-121004050826-phpapp01/85/SSONDE-Semantic-Similarity-On-liNked-Data-Entities-9-320.jpg)
![Data.cnr.it defining a context
pub:autoreCNRdi
Crawled by Data.CNR.it
dc:subject
pub: 22
skos:broader
Res 226
pub: 26 Crawled by DBPEDIA
Topic:27
Res 225 Topic:25
pub: 29
Topic:2
Topic:26
Res 226
Topic:23
PREFIX dc: <http://purl.org/dc/terms/>
No data
PREFIX pub: <http://www.cnr.it/ontology/cnr/pubblicazioni.owl#> properties are
[owl:Thing]-> {{}, { (pub:autoreCNRDi, Inter),(dc:subject, Simil)}} considered in this
context
[owl:Thing, dc:subject]-> {{},{(skos:broader, Inter)}}
Riccardo Albertoni 11](https://image.slidesharecdn.com/ssondedemotoshare-121004050826-phpapp01/85/SSONDE-Semantic-Similarity-On-liNked-Data-Entities-10-320.jpg)
![Data.cnr.it defining a context
pub:autoreCNRdi
Crawled by Data.CNR.it
dc:subject
pub: 22
skos:broader
Res 226
pub: 26 Crawled by DBPEDIA
Topic:27
Res 225 Topic:25
pub: 29
Topic:2
Topic:26
Res 226
Topic:23
PREFIX dc: <http://purl.org/dc/terms/>
PREFIX pub: <http://www.cnr.it/ontology/cnr/pubblicazioni.owl#>
[owl:Thing]->{{},{ (pub:autoreCNRDi, Inter),(dc:subject, Simil)}}
[owl:Thing, dc:subject]-> {{},{(skos:broader, Inter)}}
Riccardo Albertoni 17](https://image.slidesharecdn.com/ssondedemotoshare-121004050826-phpapp01/85/SSONDE-Semantic-Similarity-On-liNked-Data-Entities-15-320.jpg)
![What Id like you to think about
Can SSONDE be useful in any of your current
research activities ?
Can SSONDE be deployed in some of your future
projects (proposal)?
Are you interested in contributing somehow to
SSONDE?
Further details are available in
R. Albertoni, M. De Martino, SSONDE: Semantic
Similarity On liNked Data Entities, 6th Metadata and
Semantics Research Conference, 28-30 November
2012 - C叩diz (Spain) [to appear]
Framework documentation
http://code.google.com/p/ssonde/wiki/GettingStarted
Riccardo Albertoni 19](https://image.slidesharecdn.com/ssondedemotoshare-121004050826-phpapp01/85/SSONDE-Semantic-Similarity-On-liNked-Data-Entities-17-320.jpg)
SSONDE: Semantic Similarity On liNked Data Entities
- 1. SSONDE:Semantic Similarity On liNked Data Entities (Example/DEMO) Riccardo Albertoni ralbertoni@delicias.dia.fi.upm.es Ontology Engineering Group Departamento de Inteligencia Artificial Facultad de Inform叩tica Universidad Polit辿cnica de Madrid Acknowledgment: SSONDE in its current (pre)release results mainly from the research activity I did at IMATI-CNR in collaboration with M. De Martino
- 2. Linked data Crawling architectural pattern Build analysis Cluster analysis Explorative search services on resources SSONDE LDSPIDER/FUSE LDIF KI Riccardo Albertoni 2
- 3. Getting started SSONDEs code is hosted as an Google Code project, http://code.google.com/p/ssonde/ licenced as open source code (GNU GPL v3) Info about how getting started are available at http://code.google.com/p/ssonde/wiki/GettingStarted Riccardo Albertoni 3
- 4. Applying SSONDE Lets apply SSONDE on Linked data resources exposed by third parties Info about CNR provided by Gangemi et all, Data.cnr.it To work out similarities among researchers according to their research interests Date: 15/03/2012 4
- 5. What to download/install for this Example - SSONDE - svn checkout http://ssonde.googlecode.com/svn/SSONDEv1 SSONDEV1DEMO For crawling data - LDSpider - code.google.com/p/ldspider/ - Fuseski - http://jena.apache.org/documentation/serving_data/index.ht ml Riccardo Albertoni 5
- 6. Crawling RDF fragments from data.cnr.it Which resources are you interested on? 1. Identify a set of seeds entities E.g. researchers, http://www.cnr.it/ontology/cnr/individuo/unitaDiPersonaleEste rno/ID226 (Riccardo Albertoni) 2. Figure out which entity features (i.e., Object properties and Data properties) are interesting Analyse schemas which are deployed to characterize that kind of resources, e.g., http://www.cnr.it/ontology/cnr/pubblicazioni.owl (prefix:pub) pub:autoreCNRDi (the papers written by the authors) dc:subject (authors scientific interest) Riccardo Albertoni 6
- 7. Crawling RDF fragments from data.cnr.it 3. Crawl metadata of resources to be analyzed creating a seedDATACNRIT.txt file java -jar ./LDSpider/ldspider-1.1e.jar -s seedDATACNRIT.txt -b 5 -1 -oe http://localhost:3030/XXX/update 3 bis. if you have a more precise idea of which feature you need, the list of properties to be traversed can be specified by f option java -jar ./LDSpider/ldspider-1.1e.jar -b 5 -1 -f http://www.cnr.it/ontology/cnr/pubblicazioni.owl#autoreCNRDi -f http://www.cnr.it/ontology/cnr/pubblicazioni.owl#coautore -f http://purl.org/dc/terms/subject -f http://www.w3.org/2004/02/skos/core#broader -s seedDATACNRIT.txt -oe "http://localhost:3030/XXX/update" Riccardo Albertoni 7
- 8. Configuration Similarity Output List of Instances Context Layer Java Class to Similarity matrix in Ontology Layer CVS generate the list SSONDE Data Layer Ref. Context n-most similar entities Kind of Store Data wrappers In JSON Ref. Rules JENA JENA JENA Virtuoso ... (e.g., JENA rules) MEM SDB TDB Wrppr Local Data Store /Cache RDF SDB TDB Dumps virtuoso . Rep. Rep. Crawling architectural pattern WEB OF Linked data consumption DATA LDSpider +Fuseki LDIF Third parties RDF HTTP DEREFERENCIABLE SPARQL Served Linked dataset Dumps URIs End Points 9
- 9. Configuration file 1 { "StoreConfiguration":{ "KindOfStore":"JENATDB", "RDFDocumentURIs":[ ], "TDBDirectory":"data/CNRIT/TDB-0.8.9/CNRR/" }, List of LOD Entities URI "InstanceConfiguration":{ Java class Implementing ListOfInputInstances "InstanceURIsClass":"application.dataCNRIt.GetResearcherIMATIplusCoauthor" }, "OutputConfiguration":{ Similarity Matrix CSV - JSON encoding of top n-most "KindOfOutput":"JSONOrderedResult", similar "NumberOfOrderedResult":20", "FilePath":"conf/dataCNRIt/ComplexContextResearchInterest/CRRIIntPub.res.json" }, Context Encoded in a format in-house text format/ hopefully soon in JSON "ContextConfiguration":{ "ContextFilePath":"conf/dataCNRIt/ComplexContextResearchInterest/CCRIIntPub.ctx" } } Riccardo Albertoni 10
- 10. Data.cnr.it defining a context pub:autoreCNRdi Crawled by Data.CNR.it dc:subject pub: 22 skos:broader Res 226 pub: 26 Crawled by DBPEDIA Topic:27 Res 225 Topic:25 pub: 29 Topic:2 Topic:26 Res 226 Topic:23 PREFIX dc: <http://purl.org/dc/terms/> No data PREFIX pub: <http://www.cnr.it/ontology/cnr/pubblicazioni.owl#> properties are [owl:Thing]-> {{}, { (pub:autoreCNRDi, Inter),(dc:subject, Simil)}} considered in this context [owl:Thing, dc:subject]-> {{},{(skos:broader, Inter)}} Riccardo Albertoni 11
- 11. run SSONDE From the directory where you have downloaded SONDE ( e.g., SSONDEV1DEMO) java -classpath ./lib/*:./bin SSONDEv1.SemSim conf JSONconfigurationFile1 java -classpath ./lib/*:./bin/ SSONDEv1.SemSim -conf conf/dataCNRIt/ComplexContextResearchInterest/CCRIInt Pub.param.json Riccardo Albertoni 12
- 12. Ex. Of Tool to interpret SSONDE results SSONDE does not have its own GUI, results are interpreted by elaborating its similarity matrix via 1) Excel files 2) Hierarchical Clustering Explorer, 3.0, Human- Computer Interaction Lab University of Maryland. http://www.cs.umd.edu/hcil/ multi-cluster/. 14
- 13. Riccardo Albertoni 15
- 15. Data.cnr.it defining a context pub:autoreCNRdi Crawled by Data.CNR.it dc:subject pub: 22 skos:broader Res 226 pub: 26 Crawled by DBPEDIA Topic:27 Res 225 Topic:25 pub: 29 Topic:2 Topic:26 Res 226 Topic:23 PREFIX dc: <http://purl.org/dc/terms/> PREFIX pub: <http://www.cnr.it/ontology/cnr/pubblicazioni.owl#> [owl:Thing]->{{},{ (pub:autoreCNRDi, Inter),(dc:subject, Simil)}} [owl:Thing, dc:subject]-> {{},{(skos:broader, Inter)}} Riccardo Albertoni 17
- 16. What next? (i) semantic similarity optimization: (i) the caching of intermediate similarity results (ii) the adoption of MapReduce paradigm to speed up the assessment of semantic similarity; (ii) domain driven extensions at data layer: (i) defining new data layer measures suited for geo- referenced entities (ii) the multilingual similarity recently proposed by Kartic and Jorge could be considered for inclusion in the data layer of SSONDE (iii) definition of interfaces sifting entities according to their similarity, e.g., exploiting existing visualization frameworks (e.g., Exibit, Google visualization and JavaScript InfoVis Toolkit). Riccardo Albertoni 18
- 17. What Id like you to think about Can SSONDE be useful in any of your current research activities ? Can SSONDE be deployed in some of your future projects (proposal)? Are you interested in contributing somehow to SSONDE? Further details are available in R. Albertoni, M. De Martino, SSONDE: Semantic Similarity On liNked Data Entities, 6th Metadata and Semantics Research Conference, 28-30 November 2012 - C叩diz (Spain) [to appear] Framework documentation http://code.google.com/p/ssonde/wiki/GettingStarted Riccardo Albertoni 19