Query Expansion with Highly Ranked Terms
The focus on Information Retrieval (IR) has become very prominent since the arrival of the internet and finding
relevant information in a vast, and growing amount of data has become very important. One of the main goals of an IR
system is to find relevant information as accurate and efficient a process as possible. We achieve this goal by improving
search engines performance. One major problem with formulating a query is that many times user’s original query is not
sufficient to retrieve the information. Query expansion is an answer to this concern where the system adds more terms to the
user's original query to retrieve more relevant documents. Some query expansion methods have been proposed in
information retrieval. Many are based on collecting the top terms within the documents with an improvement of 10%. In this
paper, we devise a new query expansion technique called, Query Expansion using Highly Ranked Terms (QEHRT) that adds
terms to the query submitted by the user from the retrieved documents relevant to the query. It is a sequential process
composed of three phases. The experiments are done using three test collections: Medline composed of 1033 documents,
LISA composed of 5872 documents, and NPL composed of 11429 documents. We enhanced the retrieval quality of query
results by more than 15%.
Keywords - Information Retrieval, Query Expansion, Vector Model.