Abstract:
A large number of text documents have been found on the web for a different subject that
contains specific information of a particular concept. These documents are in the form of
heterogeneous nature that make the difficult task to find information from there that user need.
Now it becomes a challenging task to find relevant information from this type of source. In
order to solve this problem to get relevant information according to the user query, we
introduce an approach in the information retrieval field which is based on ontology. Here we
use ontology concept for knowledge representation in our model which organize information
in such a way that helps improve retrieval task. This technique not only helps us to solve
the heterogeneous problem of information source but also help us to retrieve semantically related
information that is the user's expectation. In this approach, we divide the documents into
different classes according to their concept matching with a class. Then we find out
contextually related feature words from each document which give us metadata for these
documents and used for indexing. So when a user asked a query, first of all, find out the query's
class then matching the query with the document's metadata, and the highest-scoring document
shows it to the user. Our approach has been tested by agriculture domain ontology which gives
us a better result.
