Enhanced Arabic Document Retrieval Using Optimized Query Paraphrasing

Query paraphrasing aims to construct a better formulation of user queries in order to enhance retrieval. Formulating search queries remains complicated for a subset of Web users. In a typical situation, a user will not receive satisfactory results from the submitted search query and will subsequently attempt different query paraphrases. The Arabic vocabulary is rich in synonyms and hyponyms. Such richness of synonyms makes automation of the paraphrasing technique crucial for Arabic information retrieval systems in order to facilitate the process of paraphrasing synonyms. In this article, we propose an enhancement for Arabic information retrieval using a query paraphrasing technique. Furthermore, two query paraphrasing optimization techniques are proposed to overcome the time complexity and exhaustive calculation of existing query paraphrasing techniques. One of these techniques uses a genetic algorithm (GA-QP), and the other employs the artificial bee colony algorithm (ABC-QP). The performance of these two algorithms is compared. ABC-QP shows an improvement in Arabic information retrieval performance compared with the genetic algorithm query paraphrasing system.