2025/01/10 by Tog‘ayev, Ilxom
paper · doi:10.34920/icdgpdt.alkhws43
this article studies the issue of computer processing of the transcription model of Uzbek phonemes. The problem is that existing transcription models cannot fully reflect the phonemic features of the language and their use in computer processing is limited. The study analyzed modern algorithmic approaches to phoneme transcription and developed a new model adapted for the Uzbek language. During the data processing, a phonetic database was created and their classification was carried out using neural networks. As a result of the study, it was found that the accuracy of the model reached 95%, which is a significant improvement over previous model. This approach opens up new opportunities for application at the intersection of linguistics and technical sciences. In conclusion, recommendations are given for the future improvement of the model and the creation of new technologies based on the phonemic features of the Uzbek language.