IEEE 14th Signal Processing and Communications Applications, Antalya, Turkey, 16 - 19 April 2006, pp.686-688
In order to increase the intelligibility and naturalness of synthetic speech, the prosodic entities such as segmental duration, pitch, and energy should be modelled appropriately. This study presents determination of the optimal feature subset for modeling segmental duration in Turkish using Classification and Regression Trees (CART) based on feature selection methods.