ABSTRACT

Yoruba language is gradually going into extinction because most speakers don't know how to write it despite that it is being taught in Primary and Secondary schools in Nigeria. This therefore call for the need of modern day processing tools such as machine translators for the language to catch up with the technological growth the world is experiencing. ln the face of rapid globalization, the significance of Machine translation cannot be overemphasized because it can translate the content quickly and provides quality output, thus saving human the stress and time of poring on translating books or looking for human translator. Hence, this research developed an Adjectival phrase-based (AD.IP) system for English to Yori1bit Machine Translation, The data for the developed Adjectival phrase-based (AD.IP) system was extracted from locally spoken words and stored in a database. The phrases were broken down into their part of speech (POS) and the database was designed by categorizing all the parts of speech into their different grammatical functions. The corpus was trained to understand the grammatical rules of translation while NLTK parser was used to parse the corpus and test all the rules used as it affects each sentence. Python programming language is the core pl'Ogramming language used in developing the system, The developed ADJP system was evaluated using human judgement by administering questionnaires to ten respondents. Expert's trans 1 ated phrases were compared to that gotten from the developed system and the respondents' using the mean opinion score (MOS) technique based on word orthography. Results show that the expert's score was I 00 percent while that of the respondents was 76.3 percent and the developed nrnchine translator value was 95.5 percent. The developed system's correctness is close to that of the Expert, and more accurate than that of the respondents giving accurate translations with appropriate tone-marks and underdots.