Investigating Vowel Reduction in Uyghur Through Machine Learning Analysis of Acoustic Features

This study explores the acoustic features of vowel production in the Uyghur language, focusing on vowel reduction in various consonantal contexts. Utilizing gradient tree boosting algorithms, we statistically model the relationship between acoustic properties and vowel qualities for thirty-two native Uyghur speakers. Our findings indicate a significant degree of vowel reduction, suggesting an ongoing sound change in the language.