Grapheme to Phoneme Conversion for Standard Malay

• Main Authors: Teoh, Boon Seong, Tan, Yeow Kee, Li, Haizhou
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2004
• Subjects: Technology and Innovation
• Online Access: https://ink.library.smu.edu.sg/lkcsb_research_smu/25; https://ink.library.smu.edu.sg/cgi/viewcontent.cgi?article=1024&context=lkcsb_research_smu
• Institution: Singapore Management University

This paper presents the use of Joint Source-Channel model (JSC) to carry out grapheme-to-phonetic (G2P) transcription process on Standard Malay (SM) [1]. Previous work on using the JSC for English to Chinese name transliteration indicates good results. Hence it is assumed that similar result can be achieved for the task of transforming SM Grapheme to SM Phoneme, especially out-of-vocabulary (OOV) SM words. This paper will discuss the SM language and the rules for text preprocessing, which are defined by [2] for SM language. A cross validation experiment was carried out and the result shows that the proposed JSC achieves an accuracy of 86.3% for the first best choice in close test and 85.7% in open test.