Key component of Microsoft Speech Server responsible for providing a SALT runtime environment for telephony applications.
For Microsoft Speech Server SDK, speech-based applications in which the user interacts with the application using voice commands over a standard telephone or a cellular phone.
A set of historical data processed by a trained data-mining algorithm. The results of processing the testing dataset can be used to determine whether the mining model was defined accurately.
A value that once met or exceeded will indicate whether a certain condition has been met. A threshold value is used in Chapter 7 to determine when a question entered by the user matches an existing question in the database.
TIM Telephony Interface Manager
A third-party component available from Microsoft partners such as Intel and Intervoice. Used by Microsoft Speech Server to interpret signals from a telephone to a telephony card..
Data-analysis technique useful in measuring sequential and non-random data sequences. A time series algorithm is expected to be released in the next version of Microsoft SQL Server.
In Microsoft Analysis Services, a set of data used to initially process a data-mining algorithm. It should contain historical data that has been cleaned to remove outliers or incomplete data.
In Microsoft Speech Server SDK, the word and phrases spoken by prompts. Can be associated with a recorded wave file imported or recorded by the SDK.
Software applications that perform most like an intelligent human being. Also commonly referred to as "Strong AI" or "Real AI."
Also known as speech synthesis. Process whereby text typed into a computer is read aloud using a computer-generated voice. For Microsoft Speech Server, this is handled by both the prompt engine and the TTS engine. The prompt engine puts together prerecorded words from the prompt database. The TTS engine produces the synthesized output heard by the user.