You are on page 1of 12

Ontology based information

retrieval model for semantic


web
Sheik sajid (10mse1104)
Mentor:Prof.Shri Devi

Motivation
The Current web-browsers and search engines
are based on the web 2.0. The companies like
Google(Freebase,Rdfa,Snippets) and facebook(Open Graph) had added semantics to the
data on the web and developing the smart
applications.
The Data on the web is Linked Data and it needs
to be analyzed rather than the unstructured one ,
query results should be distinct.

Problem statement
When a user queries about some data, the
machine interpretation of the data is necessary.
The machine should have knowledge about the
user query.
Web engines must also act as the personal
assistant and smart machines have to be
developed.
Web engines are still using the text based search
to query in complex queries.

Text search leads to hit the irrelevant pages and


retrieve results which arent useful to the user
request and To increase the precision and recall
semantic web is needed.
Semantics means adding context to the data on
the web. Information is categorized and stored
in such away that machine can understand as
human.

Key concepts in building semantic web


RDF,RDFS(Resource Description framework).
OWL-ontology vocabulary layer machine process
able semantics.
SPARQL(Query language).
Triple-stores

Architectural Design

Methodology
For Knowledge gathering protg tool is used
and it helps in reasoning
Apache jena is useful building RDF-nodes and
fuseki-server is used for querying the semantic
data.
A sparql query engine is developed by ARC2PHP based library.

The semantic search is based on the triple


stores(Subject,Predicate,Object) so the
parameters are defined in the similar way.
Retrieve the results in the desired format.
Display the relevant information in presentable
format

Modules
Building ontology's in eclipse and protg
reasoner is used to identify the unknown
knowledge
RDF parsers/serializers used for sparql queries
Website for Executing sparql query.
Get the user inputs and perform semantic search
and display the results in presentable format

You might also like