Leveraging Linked Data to Infer Semantic Relations within Structured Sources
•
4 likes•1,305 views
Information sources such as spreadsheets and databases contain a vast amount of structured data. Understanding the semantics of this information is essential to automate searching and integrating it. Semantic models capture the intended meaning of data sources by mapping them to the concepts and relationships defined by a domain ontology. Most of the effort to automatically build semantic models is focused on labeling the data fields with ontology classes and/or properties, e.g., annotating the first column of a table with dbpedia:Person and the second one with dbpedia:Film. However, a precise semantic model needs to explicitly represent the relationships too, e.g., stating that dbpedia:director is the relation between the first and second column. In this paper, we present a novel approach that leverages the small graph patterns occurring in the Linked Open Data (LOD) to automatically infer the semantic relations within a given data source assuming that the source attributes are already annotated with semantic labels. We evaluated our approach on a dataset of museum sources using the linked data published by Smithsonian American Art Museum as background knowledge. Mining only patterns of length one and two, our method achieves an average precision of 78% and recall of 70% in inferring the relationships included in the semantic models associated with data sources.
Recommended
Recommended
More Related Content
Similar to Leveraging Linked Data to Infer Semantic Relations within Structured Sources
Similar to Leveraging Linked Data to Infer Semantic Relations within Structured Sources (18)
Recently uploaded
Recently uploaded (20)
Leveraging Linked Data to Infer Semantic Relations within Structured Sources
- 1. Leveraging Linked Data to Infer Semantic Relations within Structured Sources Mohsen Taheriyan Craig A. Knoblock Pedro Szekely Jose Luis Ambite Yinyi Chen
- 2. Problem: How to map structured data to a domain ontology?
- 3. Semantic Model Map the source to the classes & properties in an ontology 3 title date name 1 The Island 2009 Walton Ford 2 Excavation at Night 1908 George Wesley Bellows 3 Rose Garden 1901 Maria Oakey Dewing SourceDomainOntology CIDOC-CRM 85 classes 297 properties
- 4. Semantic Types E35_Title E82_Actor_Appellation rdfs:label rdfs:label 4 E52_Time-Span title date name 1 The Island 2009 Walton Ford 2 Excavation at Night 1908 George Wesley Bellows 3 Rose Garden 1901 Maria Oakey Dewing P82_at_some_time_within
- 5. Relationships E35_Title E82_Actor_Appellation rdfs:label rdfs:label 5 E52_Time-Span title date name 1 The Island 2009 Walton Ford 2 Excavation at Night 1908 George Wesley Bellows 3 Rose Garden 1901 Maria Oakey Dewing P82_at_some_time_within E22_Man-Made_Object E12_Production E21_Person P102_has_title P108_was_produced_by P4_has_time-span P14_carried_out_by P131_is_identified_by
- 6. Idea • There is a huge amount of linked data available in many domains (RDF format) 6 • Use LOD as the background knowledge • Exploit the relationships between instances
- 7. Approach 7 Extract graph patterns from the linked data • Target source (S) • Domain Ontologies (O) • Semantic labels of S • Linked Data (in the same domain) Construct a graph from LOD patterns and the ontology Generate and rank semantic models 1 2 3 Input A ranked set of semantic models for S Output
- 8. Approach 8 Extract graph patterns from the linked data • Target source (S) • Domain Ontologies (O) • Semantic labels of S • Linked Data (in the same domain) Construct a graph from LOD patterns and the ontology Generate and rank semantic models 1 2 3 Input A ranked set of semantic models for S Output
- 9. LOD Patterns 9 ../person- institution/57551 E21_Person rdf:type Thomas Burgon skos:prefLabel ../person- institution/57551/birthP98i_was_born ../person- institution/57551/birth/ date P4_has_time-span E67_Birth rdf:type 1787 rdfs:label E52_Time-Span rdf:type LOD fragment from the British Museum
- 10. LOD Patterns 10 E67_BirthE21_Person P98i_was_born ../person- institution/57551 E21_Person rdf:type Thomas Burgon skos:prefLabel ../person- institution/57551/birthP98i_was_born ../person- institution/57551/birth/ date P4_has_time-span E67_Birth rdf:type 1787 rdfs:label E52_Time-Span P4_has_time-span E52_Time-Span rdf:type LOD fragment from the British Museum Pattern
- 11. Pattern Templates • Many possible templates for patterns – Example: patterns for classes C1, C2, C3 • Consider only tree patterns • Limit the length of the patterns 11
- 12. Extracting Patterns • Use SPARQL to query RDF data • Example: patterns with length 1 12 SELECT DISTINCT ?c1 ?p ?c2 (COUNT(*) as ?count) WHERE { ?x ?p ?y. ?x rdf:type ?c1. ?y rdf:type ?c2. FILTER (?x != ?y).} GROUP BY ?c1 ?p ?c2 ORDER BY DESC(?count);
- 13. Approach 13 Extract graph patterns from the linked data • Target source (S) • Domain Ontologies (O) • Semantic labels of S • Linked Data (in the same domain) Construct a graph from LOD patterns and the ontology Generate and rank semantic models 1 2 3 Input A ranked set of semantic models for S Output
- 14. Merge the Patterns into a Graph 14 E12_ProductionE53_Title P108i_was_produced_by E52_Time-Span E82_Actor_Appellation E22_Man-Made_Object E21_Person P102_has_title P14_carried_out_by P131_is_identified_by E67_Birth P98i_was_born P4_has_time-span P4_has_time-span Links are weighted: less weight for more frequent links Links have tags: the identifier of the patterns containing the link
- 15. 15 E12_ProductionE53_Title P108i_was_produced_by E52_Time-Span E82_Actor_Appellation E22_Man-Made_Object E21_Person P102_has_title P14_carried_out_by P131_is_identified_by E67_Birth P98i_was_born P4_has_time-span E39_Actor P1_is_identified_by P1_is_identified_by P98i_was_born P14_carried_out_by P4_has_time-span Add the paths from the Ontology The links added from the patterns have much less weight compared to the links added from the ontology
- 16. Approach 16 Extract graph patterns from the linked data • Target source (S) • Domain Ontologies (O) • Semantic labels of S • Linked Data (in the same domain) Construct a graph from LOD patterns and the ontology Generate and rank semantic models 1 2 3 Input A ranked set of semantic models for S Output
- 17. Map Semantic Labels to the Graph 17 E12_ProductionE53_Title P108i_was_produced_by E52_Time-Span E82_Actor_Appellation E22_Man-Made_Object E21_Person P102_has_title P14_carried_out_by P131_is_identified_by E67_Birth P98i_was_born P4_has_time-span E39_Actor P1_is_identified_by P1_is_identified_by P98i_was_born P14_carried_out_by P4_has_time-span
- 18. Map Semantic Labels to the Graph 18 E12_ProductionE53_Title P108i_was_produced_by E52_Time-Span E82_Actor_Appellation E22_Man-Made_Object E21_Person P102_has_title P14_carried_out_by P131_is_identified_by E67_Birth P98i_was_born P4_has_time-span E39_Actor P1_is_identified_by P1_is_identified_by P98i_was_born P14_carried_out_by P4_has_time-span
- 19. Generate and Rank Semantic Models • Compute Steiner tree for the mapping – A minimal tree connecting nodes of mapping – A customization of BANKS algorithm [Bhalotia et al., 2002] • Our algorithm considers both coherence and popularity • Each tree is a candidate model • Rank the models based on coherence and cost 19
- 20. What Is Coherence? 20 PlacePerson organizer Event location p1 p2 p3 PlacePerson bornIn Place isPartOf PlacePlace bornIn Person diedIn Patterns
- 21. What Is Coherence? 21 PlacePerson organizer Event location p1 p2 p3 PlacePerson bornIn Place isPartOf PlacePlace bornIn Person diedIn PlacePerson organizer Event location 1 1 bornIn 0. 5 p2, p3 p1p1 Place isPartOf p2 1 1 diedIn p3 PatternsGraph
- 22. What Is Coherence? 22 PlacePerson organizer Event location p1 p2 p3 PlacePerson bornIn Place isPartOf PlacePlace bornIn Person diedIn PatternsGraph Labels Person Place Event PlacePerson organizer Event location 1 1 bornIn 0. 5 p2, p3 p1p1 Place isPartOf p2 1 1 diedIn p3
- 23. What Is Coherence? 23 PlacePerson organizer Event location p1 p2 p3 PlacePerson bornIn Place isPartOf PlacePlace bornIn Person diedIn PatternsGraph Labels Person Place Event PlacePerson organizer Event location 1 1 bornIn 0. 5 p2, p3 p1p1 Place isPartOf p2 1 1 diedIn p3 SteinerTrees PlacePerson organizer Event bornIn p1 p2, p3 PlacePerson location Event bornIn p1 p2, p3 PlacePerson organizer Event p1 location p1 Not minimal model but more coherent
- 25. Evaluation • Correct semantic types given • Linked data: 3,398,350 triples published by Smithsonian American Art Museum • Extracted patterns of length 1 and 2 • Compute precision and recall between learned links and correct links 25 Evaluation Dataset # sources 29 # classes in the ontologies 147 # properties in the ontologies 409 # nodes in the gold standard models 812 # links in the gold standard models 785
- 27. Gold Standard Models - Example 1 27
- 28. Gold Standard Models - Example 2 28
- 29. Results 29 background knowledge precision recall time (s) domain ontology 0.07 0.05 0.17 domain ontology + patterns of length 1 0.65 0.55 0.75 domain ontology + patterns of length 1 and 2 0.78 0.70 0.46
- 30. Related Work • Mapping databases and spreadsheets to ontologies – Mapping languages and tools (D2R, R2RML) – String similarity between column names and ontology terms • Understand semantics of Web tables – Use column headers and cell values to find the labels and relations from a database of labels and relations populated from the Web • Exploit Linked Open Data (LOD) – Link the values to the entities in LOD to find the types of the values and their relationships • Learn semantic models of structured data sources from previously modeled sources – Learn from the popular and coherent patterns in known semantic models 30
- 31. Discussion & Future Work • Automatically Infer semantic relations from LOD • Help to publish consistent RDF data • Extract longer patterns from LOD 31
Editor's Notes
- Key ingredient to automate: Source discovery, Data integration, Publish knowledge graphs
- P131 is identified by is more popular than P1 is identified by to connect instances of E82 Actor Appellation and instances of E21 Person: The property P1 is identified by describes the naming or identification of any real world item by a name or any other identifier, and P131 is identified by is a specialization of P1 is identified by that identifies a name used specifically to identify an instance of E39 Actor (superclass of E21 Person).
- Keyword Searching and Browsing in Databases using BANKS
- #data source 29 #classes in the domain ontologies 147 #properties in the domain ontologies 409 #nodes in the gold-standard models 812 #data nodes in the gold-standard models 418 #class nodes in the gold-standard models 394 #links in the gold-standard models 785 #internal links in the gold-standard models 367 68 distinct patterns with length one (two nodes and one link) 634 distinct patterns with length two (three nodes and two links)
- precision: How many of the learned relationships are correct? recall: How many of the correct relationships are learned?
- Indiana Museum of Art
- Dallas Museum of Art
- Mapping databases and spreadsheets to ontologies Mapping languages: D2R [Bizer, 2003], D2RQ [Bizer and Seaborne, 2004], R2RML [Das et al., 2012] Tools: RDOTE [Vavliakis et al., 2010], RDF123 [Han et al., 2008], XLWrap [Langegger and Woß, 2009] String similarity between column names and ontology terms [Polfliet and Ichise, 2010] Understand semantics of Web tables Use column headers and cell values to find the labels and relations from a database of labels and relations populated from the Web [Wang et al., 2012] [Limaye et al., 2010] [Venetis et al., 2011] Exploit Linked Open Data (LOD) Link the values to the entities in LOD to find the types of the values and their relationships [Muoz et al., 2013] [Mulwad et al., 2013] Semantic annotation of Web services Languages: SAWSDL [Farrell and Lausen, 2007] Tools: SWEET [Maleshkova et al., 2009] Annotate input and output parameters [Heß et al., 2003] [Lerman et al., 2006] [Saquicela e al., 2011] Learn semantic definitions of online information sources [Carman, Knoblock, 2007] Learns LAV rules from known sources Only learns descriptions that are conjunctive combinations of known descriptions Learn semantic models of structured data sources [Taheriyan et al, 2013, 2014, 2015] Learns complete semantic models from previously modeled sources
- Publish linked data Transform the data to a common vocabulary Linking entities across different datasets