WHAT KIND OF SEMATICS WE ARE USING AND WHAT

SEMANTICS WE NEED TO ACHIEVE THE VISION OF

COMPLETE SEMANTIC WEB

1 Introduction Web” suddenly put knowledge acquisition

In this age of information, search engines and knowledge representation at the centre
act as important services, providing the of an extremely interesting and powerful
community with the information hidden in research areas. So first step in semantic
the web and, due to their frequent use, search engine architecture is Semantic
stand as an integral part of our lives [1]. Knowledge extraction which is later used
The last decade has witnessed design and by semantic search engine for retrieval of
implementation of several state of the art relevant information lot of research have
search engines. Today, there are search been done but still it is at the very nascent
engines that have indexed more than four stage if we are looking for a complete
billion web pages [2], processing millions intelligent web.
of user queries per day over their local Many eminent researchers are working in
index. The Technology of searching is this direction but we have achieved a very
shifting from traditional keyword based little bit of semantic web as required
search engine to semantic search engine in Different semantic search engines are
which main emphasis is on getting the developed but they are acquiring very less
meaningful information and we are semantic knowledge from the web. So in
moving from traditional web to semantic this paper we present a survey of some
web. recent semantic search engine and what
kind of semantic knowledge they are
Although SemanticWeb research is still a extracting and the different semantic
young discipline a fair amount of research knowledge the semantic search engine
efforts are going on in this direction of the require to extract to achieve the complete
vision of [Berners-Lee et al., 2001] . A vision of semantic web. we divide this
major category in this area of semantic paper into three parts
web research is knowledge extraction
(which deals with the problem of 1. Introduction
extracting and modelling knowledge) and 2. Survey of the present semantic search
knowledge representation (which deals engines what kind of semantic knowledge
with the problem of representing they are extracting.
knowledge and reasoning about it) A lot of 3. Different kind of semantic knowledge
research has been done and going on in we require to extract to achieve the full
both these areas. vision of semantic web.
However, the results of both communities 4. Conclusion
are not up to the mark as the expectations
of the real semantic web. Knowledge 2. Survey
extraction is too costly and present This section gives an overview of
techniques are extracting very less semantic knowledge used by various
semantic knowledge as required, and the semantic search engines to search in the
knowledge representation systems that context of the Semantic Web and what
were created were mainly isolated, brittle, kind of semantic information they are
and small solutions for minor problems. extracting.
Transforming the Web into a “knowledge
MindNet:- MindNet represents a general states relations and concepts as first-order
methodology for acquiring, structuring, object.
accessing, and exploiting semantic SWE [32] is capturing context in terms of
information from natural language text. Temporal and Space attributes.
This paper provides an overview of the
distinguishing characteristics of MindNet. SHOE :- SHOE() is a Search tool, which
The semantic knowledge used by mindnet allows the user to specify a context for his
is Attribute goal possessor or her query, and then uses the context to
SECO:- an application that collects and help the user build a query by example. In
aggregates data from the Semantic Web. In SHOE, each concept is called an instance
SECO, both RDF data and data from and is assigned a key, typically the URL of
legacy data sources using wrappers are the web page that best represents the
collected, the resulting data is cleansed and concept. The user can then express that the
can be queried by software agents via a concept is a member of a particular
remote query server. Seco use Resources category or that it has certain relationships.
and their properties as a semantic These relationships may either involve
information. literal values such as strings or numbers
OntoSearch:- OntoSearch() uses concepts (as may be the case with the relations
as the seeds to semantic network based name or height) or can be with other
domain ontology, the spreading activation instances, in which case the key of the
theory process infers the concepts that are related instance must be known. Since
semantically related to the initial concept there are no guarantees about the accuracy
set. Finally, the conceptual relevance of information on the Web, category and
scores, in terms of the concept activations relation declarations such as these are
in the domain ontology, are used to re-rank referred to as claims.
the documents before presentation to the
user. OntoSearch:- OntoSearch, a full-text
Beagle++;-In this author uses a contextual search engine that exploits ontological
ontology to extract the information from knowledge for document retrieval, is
the different web documents presented in this paper. Different from
other ontology based search engines,
KIM: KIM [Kiryakov et al., 2004] relies OntoSearch does not require a user to
on information extraction and relates specify the associated concepts of his/her
words in documents with concepts from an queries. Domain ontology in OntoSearch
ontology. Before indexing, documents are is in the form of a semantic network.
enriched with identifiers for the Given a keyword based query, OntoSearch
ontological concepts the words in the infers the related concepts through a
document represent. These identifers are spreading activation process in the domain
directly inserted into the indexed text. For ontology. Concepts are represented as
homonyms the same identifier is used. nodes in the network and linked through
Queries are formed using concepts and relations.So basically in this paper Xing
relations from the ontology Jiang(et.al) uses the concepts, relationship
SEAL :- SEAL [Stojanovic et al., 2001] is as a semantic information.
a framework developing semantic
portals.Seal uses a knowledge warehouse,
i.e. the ontology and knowledge base, to
extract the relevant information. The
knowledge warehouse is organised around
a relational database, where facts and
concepts are stored in a reified format. It