The Qur'an is considered the first source of knowledge and guidance for Muslims throughout the world. It is hard to understand and interpret without consulting domain experts and specialized Qur'anic books. Therefore we believe that a system based on simple questions written in Arabic and capable of retrieving answers from the Qur'an would be of a great interest to all those who want to study the Qur'an. In recent years, a number of researches have been conducted to facilitate the retrieval of knowledge from the Qur'an; however, most of the available researches are based on keyword search and do not rely on semantics. Building a semantic-based system has a number of challenges such as the lack of resources for the Arabic language and the difficulty to model the content of the Qur'an by fear of altering its right meaning. In this paper, we introduce a semantic-based search engine for the Qur'an, it is based on creating an ontology that represents the Qur'an knowledge in Web Ontology Language format, and a natural language interface that transforms user queries expressed in Arabic into SPARQL queries and then retrieves answers from the ontology.

No abstract

With the growing expansion of the semantic web and its applications, providing natural language interfaces (NLI) to end-users becomes essential to querying RDF stores and ontologies, using simple questions expressed in natural language. Existing NLIs work mostly with the English language. There are very few attempts to develop systems supporting the Arabic language. In this paper, we propose a portable NLI to Arabic ontologies; it will transform the user's query expressed in Arabic into formal language query. The proposed system starts by a preparation phase that creates a gazetteer from the given ontology. The issued query is then processed using natural language processing (NLP) techniques to extract keywords. These keywords are mapped to the ontology entities, then a valid SPARQL query is generated based on the ontology definition and the reasoning capabilities of the Web Ontology Language (OWL). To evaluate our tool we used two different Arabic ontologies: a Qur'anic ontology and an Arabic sample of Mooney Geography dataset. The proposed system achieved 64% recall and 76% precision.