المستخلص: |
Machine Translation (MT) is one of our great dreams in computer application. Malay language is becoming essential tool for Arab people especially in Malaysia. To satisfy this need, we develop the techniques for the Arabic Malay dialogue translation system. The problems of this research is different structure of interrogative dialogue sentences, verb conjugated in dialogue sentences, and ordering of adjective and adverb in all types of dialogue sentences. This research focuses on the analysis of morphology and syntax of Arabic and Malay dialogue sentences to produce grammatical rules that can be apply on Rule Based Machine Translation approach to translate Arabic to Malay dialogue. Rule Based Machine Translation approach (RBMT) is based on certain rules to convert source text (Arabic) structure to target text (Malay) structure. It includes a source language sentence, analysis module, transfer module, and generation module. This research contains fifty dialogue sentences (50) in medical domain (dialogue between doctor and patient) have been tested by our system (computer translation algorithm) and the results were compared with human translation. The result shows around 87.2% accuracy of the dialogues translation from Arabic to Malay languages. This study shows 13% incorrect translations due to applying incorrect rules or there is ambiguous of words.
|