RJET-Abstract

Development of English to Hausa Machine Translation System


Safiriyu I. Eludiora, Beatrice O. Afolabi and Abdulrahman A. Sanusi

Review I Published September, 2016


Research Journal of Engineering and Technology  Vol. 2 (1), pp. 12-32.

 

ABSTRACT

 

The Hausa language is one of the major indigenous languages in Nigeria.  Hausa language speakers cover the Northern parts of Nigeria, West Africa and Africa in general. Hausa language used in media houses for news and in Schools for teaching.  The development of English to Hausa machine translation (E-HMT) system will assist the interested learners to learn the language. Hausa is a tone language like Yorùbá. It was observed that the tone marks and diacritics are not used in writing unlike Yorùbá that the tone marks and diacritics must be used. The grammar structure of the two languages are subject verb object (SVO), but with some differences. Findings show that Google has developed E-HMT using the data driven approach. In the study reported in this paper rule-based approach was used to develop the E-HMT system. The E-HMT system can translate simple sentences. Home domain terminologies were used for the development of E-HMT system database. Some linguistics and computing theories and rules were used for the E-HMT translation process. The re-write rules were designed for the two languages. The re-write rules were tested using JFLAP. Unified Modeling Language (UML) was used to design the software. Java programming language was used for the implementation of the software. The E-HMT system could have been evaluated using Mean opinion score, but a good number of Hausa speakers in our community are not literate. The questionnaires were administered at Kaduna. We compared the E-HMT system outputs with the Hausa Google translator. The results show that E-HMT system performs better than Hausa Google translator.

Key words: Hausa, Ruled-based, re-write rules, ambiguity and Tagging.

Full Text-PDF