Pronunciation by analogy

The decision function can take a number of forms, and finding better methods of designing this function is the main goal of research in prommciation by analogy. Marchand and Damper [300] give a review of some techniques. The original approach is to find the shortest paths through 

---

The statistieal approach has been somewhat of a latecomer to grapheme-to-phoneme conversion, perhaps because of the success of other data driven techniques such as pronunciation by analogy or the impression that context-sensitive rewrite rules are adequate so long as they can be automatically trained, e.g. by a decision tree. In recent years however a number of approaches have been developed which give a properly statistical approach. 

Many data driven techniques exist, including neural networks, pronunciation by analogy and decision trees. 

Damper, R., and Eastmond, J. Pronunciation by analogy Impact of implementational choices on performance. Language and Speech 40, 1 (1997), 1-23. 

Marc HAND, Y, and Damper, R. A multistrategy approach to improving pronunciation by analogy. Computational Linguistics 26, 2 (2000), 195-219. 

Dedina, M., and Nusbaum, H. Pronounce A program for pronunciation by analogy. Computer Speech Language (Print) 5, 1 (1991), 55-64. 

---