Hybrid approach for aligning parallel sentences for languages without a written form using standard Malay and Malay dialects

摘要

Alignment of parallel text is a step for building a machine translation. Parallel text alignment is important because linguistic information can be retrieved from the result of alignment which including bilingual dictionaries and grammars correspondence of each language. In this paper, we propose a hybrid approach for standard Malay-dialectal Malay parallel text alignment. The Malay dialects in Malaysia can be grouped according to the states such as Perak dialect, Kedah dialect and Terengganu dialect. It is important to learn Malay dialects as it is still flourished and widely used in many areas especially for unofficial matters. Kelantanese Malay is used as an example for dialectal Malay in this research. The obtained precision and recall values of the proposed alignment methods are above 90%.