Inference of letter-phoneme correspondences with pre-defined consonant and vowel patterns

This paper describes the automatic inferencing of letter-phoneme correspondences with pre-defined consonant and vowel patterns, which imply a segmentation of the word in one domain. The technique obtains the maximum likelihood (ML) alignment of the training word, and correspondences are found according to where the segmentation projects onto the ML alignment. Here, the phoneme strings were segmented depending on the number of consonant phonemes preceding or following the vowel phoneme. Sets of correspondences were evaluated according to the performance obtained when they were used for text-phonemic alignment and translation. The number of correspondences inferred was too large to evaluate using Markov statistics. Instead, hidden Markov statistics were used where the storage demand is further reduced by a recoding technique. Performance improves significantly as the number of consonants included in the pattern increases. Further, the performance of correspondences with pre-defined V.C* patterns was consistently better than those with C*.V patterns.