The increasing of interest in Arabic natural language processing and semantic Web research involves an emerging need to the development of new Question Answering Systems (QAS). These systems allow users, who ignore both the structure of data and the complicated query languages, to ask a question in Arabic natural language and get the relevant answer. However, most existing QA systems focused on English and Latin-based languages. Less effort has been concentrated on the Arabic language which belongs to "Right to Left Semitic Languages". This work is an early version of a new domain-independent Arabic question answering system over linked data, which aims to particularly help Arab users to explore the Arabic semantic Web based on Arabic ontology. We describe with sufficient details the different modules of our proposed system, which uses language parser, finite state automaton and semantic Web techniques to linguistically process and answer the Arabic natural language question. Experiments have been carried out to evaluate and show efficiency of the proposed system.
Key words: Question answering system, Natural language processing (NLP), Linked Data, Arabic language, Semantic Web, Ontology, SPARQL.
|
