الفهرس 
CHAPTER 1: IntroductionOnly 14 pages are availabe for public view
 |  | from | 142 |  |  |
| | | from | 142 | | |
Abstract
During the last decade, the growth of the internet and the
availability of enormous amounts of data in digital form have necessities
a lot of interest in techniques that help the user in retrieving the needed
data of interest. Many efforts are automatically being done nowadays to
extract and retrieve information from huge repositories. Information
Retrieval (IR) systems are frequently used for such task. Information
Retrieval is the activity of obtaining material usually textual which
satisfies user need from within large collections. User can also utilize
information retrieval to browse or filter a document collection or analyze
a collection of retrieved documents. The system looks through billions of
documents on millions of computers. Information Retrieval is considered
to be the main form of Information access.
The (IR) system assists users in finding the information they require but
it does not explicitly retrieve the answers of the questions. The user
simply requires a single piece of information instead of a list of
documents. Therefore to retrieve the exact answer to question, Question
Answering Systems are used.
Question Answering (QA) is a branch of computer science that combines
Information Retrieval and Natural Language Processing (NLP) to create
systems that automatically answer questions presented by humans in
natural language. The concept of question answering systems represents a
significant improvement in information retrieval technologies,
particularly in terms of its ability to access knowledge resources in a
natural manner by querying and retrieving appropriate answers.
In order to get more accurate and exact answer, Semantic Question
Answering Systems are used. In these systems, computers are able to
make meaningful interpretation of the user’s question as well as, the
user’s intent. It takes a user’s question in natural language, processes it,
and searches the knowledge base for the answer. Finds the most
appropriate result and delivers the answer to the user which is relevant to
his or her question.
Knowledge base is used to store structured and unstructured information.
One of the most essential knowledge base is ontology. Ontology is one of
the most significant standard knowledge representations for the Semantic
Web.
This thesis presents A Semantic Question Answering System Using
Dbpedia ontology. The proposed system was built using Dbpedia
Ontology, which is a structured version of Wikipedia that incorporates
semantic knowledge. Our model allows users to ask questions in natural
language, and the system will return accurate answers to users directly
after the analysis of the question and the extraction of the answer
depending on Dbpedia ontology. The system consists of three
components, which are Question Classification – Question Processing
– Query Formulation and Execution.
Question Classification is in any QA system, categorizes questions into
one or more classes. So it determines the answer type which facilitates
answering the question. Previous research has demonstrated that
correctly, predicting the intended answer type is critical to the overall
performance of a question answering system.
In the Question Processing stage the question is analyzed. The
question’s resources and keywords were extracted. After the extraction of
keywords from the question, the synonyms of these keywords were
extracted from a website (merriam-webster.com). Synonyms from
WordNet were also used to enrich the keywords.
In the Query Formulation and Execution stage a Sparql query was
built with the resource. Sparql (SPARQL Protocol and RDF Query
Language) is a semantic query language for databases. The query returns
an RDF file containing all of the resource’s ontology classes and
properties. Similarity between the keywords (synonyms) and ontology
properties, classes was computed where the ontology class with the
highest similarity was chosen to be utilized in the final Sparql query. The
final Sparql query is built to answer the question from Dbpedia server.
This system is tested by using a set of 400 Questions collected from
different sources manually and the results are compared with another QA
system (SELNI). It’s found that the proposed system could provide
answers for 302 questions and SELNI system could provide answers for
105 questions. The accuracy of our system is 75.5%. Finally, the
experimental data is performed to show the applicability and efficiency
of the proposed model