Statistically trained orthographic to sound models for Thai

Explore this paper's citation graph

Summary

This paper presents a letter-to-sound rule system for Thai, that is trained automatically from lexicons, that combines the n-gram of phone model with the decision trees, to achieve word accuracy better than 65.15% in the rule-based approach.

Type
article
Published
2000-10-16
Cited by
32
References
10

Keywords

Computer science, Sound (geography), Speech recognition, Orthographic projection, Artificial intelligence

References

Cited by

Related papers