Query Reformulation Using Crop Characteristic in Specific Domain Search

Abstract

Retrieving relevant information from Web is a challenging task. This is due to the size and rapid growth of the Web content. The existence of great search engine such as Google or Bing has simplified our searching mission. But those search engines are only efficient if the user submit query that precisely represent the user's search intent to the system. The fact is users are bad in translating their search intent into a query format for the search system to process. Therefore, users reformulate their query several times for getting more relevant results. The aim of this paper is to present evaluation of seven distinct techniques employed in query reformulation process. This work proposes the use of crop characteristic category element in reformulating query, as well as combination with ontology and query keyword elements. The experimental results show that the use of crop characteristic category in reformulating query has significantly improved the search results performance. In the future, it also can be used to replace the ontology element if the ontology is not available.