Definition
The process of creating statistical representations of various speech sounds and their characteristics for use in automated speech recognition systems.
In short: The process of creating statistical representations of various speech sounds and their characteristics for use in automated speech recognition systems.