Davel, MHBarnard, E2012-01-182012-01-182004-11Davel, MH and Barnard, E. 2004. Default-and-refinement approach to pronunciation prediction. 15th Annual Symposium of the Pattern Recognition Association of South Africa, Grabouw, South Africa, 25 to 26 November 2004http://hdl.handle.net/10204/550115th Annual Symposium of the Pattern Recognition Association of South Africa, Grabouw, South Africa, 25 to 26 November 2004The authors define a novel g-to-p prediction algorithm that utilises the concept of a 'default phoneme': a grapheme which is realised as a specific phoneme significantly more often than as any other phoneme. They found that this approach results in an algorithm that performs well across a range from very small to large data sets. The authors evaluated the algorithm on two benchmarked databases (Fonilex and NETtalk) and found highly competitive performance in asymptotic accuracy, initial learning speed, and model compactness.enNeural networksDecision treesPronunciationAnalogy modelsInstance based learning algorithmsDynamically expanding contextPRASA 2004Conference PresentationDavel, M., & Barnard, E. (2004). Default-and-refinement approach to pronunciation prediction. PRASA 2004. http://hdl.handle.net/10204/5501Davel, MH, and E Barnard. "Default-and-refinement approach to pronunciation prediction." (2004): http://hdl.handle.net/10204/5501Davel M, Barnard E, Default-and-refinement approach to pronunciation prediction; PRASA 2004; 2004. http://hdl.handle.net/10204/5501 .TY - Conference Presentation AU - Davel, MH AU - Barnard, E AB - The authors define a novel g-to-p prediction algorithm that utilises the concept of a 'default phoneme': a grapheme which is realised as a specific phoneme significantly more often than as any other phoneme. They found that this approach results in an algorithm that performs well across a range from very small to large data sets. The authors evaluated the algorithm on two benchmarked databases (Fonilex and NETtalk) and found highly competitive performance in asymptotic accuracy, initial learning speed, and model compactness. DA - 2004-11 DB - ResearchSpace DP - CSIR KW - Neural networks KW - Decision trees KW - Pronunciation KW - Analogy models KW - Instance based learning algorithms KW - Dynamically expanding context KW - PRASA 2004 LK - https://researchspace.csir.co.za PY - 2004 T1 - Default-and-refinement approach to pronunciation prediction TI - Default-and-refinement approach to pronunciation prediction UR - http://hdl.handle.net/10204/5501 ER -