Digital Quran is a natural language document that use either Arabic font or images of the verses. In the Al-Quran there are 18994 unique words. Thus, the image approach uses a significant amount of memory space. However there is not much work has been done using machine translation (MT) technique for the Quranic representation. This paper will proposed Arabic information retrieval based on keywords search in Hexadecimal Representation using Al-Quran verses as the test case. All Quranic words will transliterate into machine language in the form of binary format after removing diacritic and duplication. This machine language approach in representing Digital Quran reduces the size of storage around 47-54% and retrieval time up to 20% hence reduce the query cost for Arabic information retrieval in general.

Digital Quran is a common application today in smart devices such as hand phones and tablet. The verses are usually presented using images of verses on written in an Arabic font. This raise issues related to storage when representing the whole Quran text. This study proposed a hexadecimal digital representation technique for words in Arabic using UTF-8 for character encoding which is backward compatible with ASCII code. This paper will explain three approaches used in representing the digital Quran. They are the hexadecimal representation for each Quranic words, sparse matrix in representing verses of Al-Quran and double offset indexing for transforming the sparse matrix to efficiently use the space. This representation proof more than 50% reduction in the memory space for storage thus will increase the searching speed. The storage of words being optimized through the use of one memory space for that particular word rather than one memory space for each Arabic character in the words. This technique helps optimized memory management for the whole digital Quran. Due to the smaller size, the digital Quran can be conveniently installed in the future as standard applications in mobile devices.