Slide 1Graph Theory Underpins RDFRDF vs RDBMSRelational Databases and RDFExample DomainsRelational Databases and RDFSlide 7Semantic Web Vs Database FederationSlide 9Slide 10Slide 11Practical Application (continued) - With a Graph DatabaseSteven SeidaEdited: Keven L. AtesWhy Use Semantic Technologies Like RDF?Graph Theory Underpins RDFGraphs (Nodes & Links) are Used to Represent a Knowledge Domain and Specifics (Instances) of ItKnowledge Domain (Ontology)Television NetworksInstance DataBroadcastNews ProgramsNews ArticlesContainingABCABC Nightly NewsBroadcasts…Competes WithFoxContaining…RDF vs RDBMSCriteria Knowledge Enablement RDBM ApproachEvolving problem space  - Adding knowledge representations and data is easyAdding related columns is hard, adding related tables is very hardInference across problem domains and multiple sources* - Intended for inference and general connectivity analysisGenerally no real inferencing framework. Very hard to link data across domains.Simple query performance (simple joins and structured query)Moderate performance after tuning - Excellent performance after tuningSocial query performance (complicated joins and recursive query to traverse a large graph) - Moderate performance after tuningLots of software required and very brittleLots of missing data  - Absent data is no issue Empty values are a headache* For example – Finding that a suspect terrorist had a friendship in college with the newly appointed chief of a power plant.Relational Databases and RDFRDF Concept Relational Database MappingClass TableDatatype Property Field/column (with a value)Object Property Table Relationship (Related Column)Add Class Type – extend ontology (add a few triples)Add Table – potentially impact entire schemaAdd Datatype Property – extend ontologyAdd Fields (and populate or program how to handle missing values)Add Object Property – extend ontologyAdd table and columnsA strong focus of RDF is storage. So, when should RDF be used rather than Relational Databases?Answer: Whenever your problem domain is dynamic/growing.Example DomainsGeneral Domain Static Portion Dynamic PortionStock Market Stock Data (prices, quantities, types of instruments – stock, options)Ways to try to make money using the stock marketReal Estate Property Data (prices, locations, owners, agents, mortgages)Techniques for optimizing return on investmentIntelligence CommunityConventional Warfare (ground, tanks, ships, land)Unconventional Warfare (Improvised explosive devises, terrorist techniques, etc)Relational Databases and RDFA strong focus of RDF is exchange (i.e., public access store, published serialized data). When should it be used?Answer: When you need to federate multiple data stores, especially those containing heterogeneous information.“There is no standard serialization of a relational database that will allow it to be imported into a competing relational database system without a change of semantics.” Allemang and Hendler – 2008Semantic Web for the Working OntologistReplace ETL with Semantic Layers*ETL – Extract, Transform, Load – database transfer schemeETL ReplacementSemantic Web vs Database FederationDatabaseQuery EngineApplicationAnalyticsUser Interface…DatabaseIdentical Structures and Duplication Required!RDF Store(merge)Query EngineTypical Database FederationSemantic Web FederationApplicationAnalyticsUser Interface…Webpages, Spreadsheets, Tables, Databases, etc.Converters and ScrapersRDF Stores and Files-No Duplication...or Federated Data StoresPractical ExamplePractical Application -An Artist’s Impression of Knowledge in a RDBMSRose Kennedy’s FamilyPractical Application (continued) - Knowledge in Triples (RDF) Form…32: "person2” "type” "person”33: "person2” "first-name” "Rose”34: "person2” "middle-initial" "Elizabeth”35: "person2" "last-name" "Fitzgerald”36: "person2” "suffix” "none”37: "person2” "alma-mater” "Sacred-Heart-Convent”38: "person2” "birth-year” "1890”39: "person2” "death-year” "1995”40: "person2” "sex” "female”41: "person2” "spouse” "person1”58: "person2” "has-child” "person17”56: "person2" "has-child” "person15”54: "person2” "has-child” "person14”52: "person2” "has-child” "person13”50: "person2” "has-child” "person9”48: "person2” "has-child" "person7”46: "person2” "has-child" "person6”44: "person2” "has-child" "person4”42: "person2" "has-child" "person3”60: "person2” "profession" "home-maker”…Practical Application (continued) -With a Graph Database•You add new predicates without changing any schema•One-to-many relations are directly encoded without the indirection of tables•You never think about what to index because all the predicates are