Construction of < Knowledge Base > 6-1 Semantic Web Semantic

Summary: About the composition of URIs, namespaces, namespaces Prefix,curie,turtle, about RDFA

Semantic Web semantic: The goal is to build a web so that all data has a standard format

RDF Resource Description Framework: is a kb

Based on: entity,classes, two yuan relationship, labels

Yago: is KB, the purpose is to automatically extract useful information from Wikipedia or other websites

Nell:language Learner Automatic Information extraction

DBpedia: Same as the yago purpose

Wikidata: Create KB so that data can be modified by machines and people

BabelNet: Establishing a multi-language ontology multilingual

About URIs

Namespace namespace: A collection of names

For example: the name of the namespace: KB1, the namespace contains: elvis,baba,etc

Local name: the names contained within the namespace

Qualified name qualified name: Name of namespace + local name

Example: Kb1:elvis

URI Unified Resource Identifier:

For example:

All URLs are URIs, but not all URIs are URLs.

About namespaces: http://elvis.is/king/of/sing

Http://elvis.is is the WW mapping to domain owner.

The rest is the responsibility of the domain owner.

Overlapping words are not allowed

URI is completely ambiguity

You can synonym, a thing can be pointed to by multiple URIs, but the same URI points to only one thing

The first part of namespace Prefix:uri for example @prefix dbp:www.dbpedia.org

Curie/qname:prefix's name +localname For example Dbp:elvis represents Www.dbpedia.org/Elvis

The first part of base Uri:uri

Turtle: A special notation for RDF

For example

@prefix P:<uri>.

@base <uri>.

Or define a type of litteral

such as "literal"?? DataType Here The datatype is given by URI or Curie

Or: Y:elvis y:loves Y:priscilla.

KB can describe the relationship of an entity between different kilobytes, such as using turtle

For example: @prefix p:<uri>.

@base <uri>.

KB can define an entity in another KB

Deferenceable uri/cool URI: The URI is accessed by the RDF client, and the fragment of the URI is returned

KB Alignment: Maps an instance label in one KB, and so on, into another KB

Algorithm:

-Literrals Matching

-assuming that their corresponding relation are matched

n Local functionality:1/y:r (x, y) an X has a number of Y values, the bigger the better.

n Inverse functionality:1/x:r (x, y) a Y is the number of X link values the bigger the better

n functionnality:x the number of/xy between the number of connections to Y

-If the inverse functionality is particularly high, the match is calculated

-If the SBJ and obj of the overlap of the two relation are very many, the match is counted.

-Calculate Subsumption's class

-Cycle

Linked Data: Let the Machine access kb

RDFa: Commenting HTML into RDF

-Vocab to define the local name in the HTML node

-Ressource defines the SUBJ in HTML

-typeof defines the type of SUBJ in HTML

-property defines the URI of the HTML subj

-Property+typeof define a new entity

RDFA Benefits: No duplication, publisher is independent

JSON Based Rdf:json-formatted RDF

The syntax of RDF A-rdf:type->b

Dublincore: KB of entities describing the file

For example: Dc:creator, Dc:title, Dc:format,

Schema.org:KB

Creative COMMONS:RDF

Everyone can build URIs

URIs do not make sense to do distinct, it is disambiguous in itself

Construction of < Knowledge Base > 6-1 Semantic Web Semantic