Moreover, the simple and open nature of RDF can create any triple. Agents or machines cannot know if author can be a machine or thing or person. RDF Schema when imported to the model and merged with RDF triples, adds some inferencing and semantics to the basic RDF.
RDFS is thus adds some semantics to RDF by defining classes, subclasses, properties and subproperties. RDFS enriches the description of existing triple statements.
All this is good but although schema can serve as a small ontology for machines to inference the RDF data, but higher level semantics cannot be defined just with Schema. For instance cardinality of properties or relationships, mapping of similar properties cannot be defined with schema. (A detailed example below).
Web Ontology Language(OWL), an extension for RDF, the next layer in the semantic web cake helps in that aspect. It can help in designing ontologies with constraints such as cardinality and datatyping.
OWL was started to help softwares to act autonomously without humans updating the software as in other data formats and models like relational database and UML. It is the next important thing after RDF in the Semantic Web Technologies as it helps in defining unambiguous, complex and interdependent data models that rely on mathematical logic.
What it does?
Owl is a declarative language that is used in expressing ontologies
OWL along with DL usually can help defining the terminology and describing classes, subclasses, relationships in a particular domain and thus create Ontologies.
Basically, it helps in further adding semantics, reasoning of inferred triple statements from Schema.
What is the need for Ontology?
Let’s look at a practical real-world example:
Although we are talking about RDF model, the practical world is full of relational databases.
RDF thrives on using URIs to represent information instead of using words (the process of removing ambiguity and assuring certainty)
For example, let’s consider collecting the address of working people in a particular zip code using the databases of companies in the city.
When it is human doing the task of filtering and collect, we can understand the meaning of the word of database fields. (even if they are named with synonyms)
But with machines, they cannot identify which field in every database is zip code. Some databases have it as address code, some as pin codes and some might be as area codes.
A software agent designed to combine databases and filter the results from the combination should have an easy way to understand which field in different databases defines zip code
To help such situations, ontologies, the other important and basic component of semantic web is used. Ontology of a domain defines such similarities and relate the terms accordingly.
Ontologies in terms of philosophy is the study of existence, the details of what type of things exists and their theories.
Considering the similarities, web and related fields, named collection of terms and the relations between those terms as ontology.
Practically in web related terms, Ontology is a document or a file that formally defines the relation between terms using taxonomies and set of inference rules.
Taxonomies: Ontologies has class and sub classes of objects(entities) and relationships amongst them(classes and subclasses)
For example we ourselves use categories and sub categories on our blogs. The categorisation and relation amongst categories helps in managing our content and gives some meaning to the sitemap of the website.
Let’s assume web is one giant website(a directory of all websites)
Such basic classes and subclasses of web entities will become very powerful tool (in terms of management and traversal)
Inference rules in Ontologies
Besides categorisation, the other thing that empowers ontologies is the inference rules. They formally define what category is a subset of other category and what all properties of the mother category are inherited by child category by default.
For example, here is what inference rules in ontology can do
By intuition, we know, a neighbourhood belongs to state, state belongs to a country. Ontologies define such things for machines.
Software agents can use such ontologies to readily suggest us ( with more certainty as well than the information we get on the search engines), for instance the date format associated with a particular neighbourhood (as the agent )
Besides adding better semantics, ontologies provide more advantages.
Accuracy as result
Accuracy is one of the advantages of semantic web we have been discussing in this paper.
Ontology part of the semantic web cake helps in providing that accuracy by removing the ambiguity.
For instance, web pages with a link to a relevant ontology of their niche, is already explaining the software agents of the web that, this page is about a certain topic.
For the growth of knowledge base: When websites link their pages with ontologies and define their entities with ontology rules, the knowledge on the web page can attributed to the knowledge base associated with Ontology used.
RIF can be used for those ontologies with complex relations and concepts where OWL is not an efficient idea to handle the complexity.
RIF: After owl, rules are the next important thing.
Different applications on the web use their own rule systems.
In the same way, ontologies of different domains with different rules will be used together on the semantic web.
For instance there are ontologies about auto parts of different car companies. One of those companies might use measurements in inches and other might use measurements in centimetres.
For semantic web to make sense of both in the process of inferencing about auto industry, need a mechanism to translate the rules of one ontology with the other. RIF, as the name specifies, rule interchange format serves this purpose.
For instances there are Rule systems like FLORA and Euler for decision making in different health care systems. When those two systems have to integrate, systems like RIF are essential for data interoperability.
One of the goals on the semantic web is “reasoning of the information on the web”
RIF aims at such reasoning by implementing the seamless rule interchange between rule languages in semantic web
Rules on the semantic web are there generally to enhance expressiveness in ontologies.
SPARQL: Just like SQL is being used to query the relational databases, SPARQL is used to query the RDF based data documents.
It is the layer to unify the logic by using mathematical logic to reconcile all semantics of other parts RDF, RDFS, OWL, SPARQL and RIF into a consistent model.
It would be complicated for applications to follow the rules of all the semantic technologies while integrating with other applications. One unified logic would make it easy for application developers to develop compatible models.
Is the information provided by semantic web is right after all?
There are a few layers which provides new inferences and rules on the semantic web layer cake which will further lead to some conclusions or more new inferences. . Proof layer on the top let the user find out which business rule or inference has led to the conclusions.
All these technologies are developed with an idea of enabling machines understand the context of the webpages to automate things associated with web.
This can happen with intelligent agents and software. agents. Let’s dive into the concept of agents.
The power of semantic web lies in Software agents:
The present value of the web on internet is so high. There a multitude of reasons for that but here are a few reasons that created this immense value
People create websites
People use links to point to other website in the effort to create a better value.
In the same way for semantic web can become as powerful when people create web agents to collect information from diverse sources, process it and exchange the data and results in between.