Knowledge representation model for inference of English text-to-phoneme conversion

One of the problems encountered in speech synthesis and electronic dictionary is that of associating with each of the phonemic realizations. To deal with this problem, some methods have been given by many other people. All these methods employ various kinds of rules and corresponding rule interpreters. But the rules involved have not a simple, unified formation. So the algorithm is difficult to handle with. A text-to-phoneme component has been developed which employs: (1) a knowledge representation for English word phoneme, (2) a rule base corresponding to this representation, (3) an interpreter for applying this representation to the rules. Some details are described here including the basic forms of the knowledge representation, the algorithm of the interpreter and three types of rules