A METHODOLOGY FOR DIAGNOSTIC EVALUATION OF SPOKEN HUMAN-MACHINE DIALOGUE

Diagnostic evaluation is an important instrument for the development o f high-quality spoken language dialogue systems. Yet no rigorous metho dology exists for the systematic and exhaustive diagnostic evaluation of all aspects of spoken language interaction: recognition, synthesis, grammar, vocabulary, dialogue, etc. The paper addresses part of this problem by presenting a methodology for the diagnostic evaluation of s poken human-machine dialogue. The first part of the methodology is gen eral and supports the detection of any kind of user-system interaction problem. The second part concerns the classification, diagnosis and r epair of problems of dialogue interaction. These problems are of two b road kinds: dialogue design errors and user Errors. Use of the methodo logy is demonstrated through analysis of a corpus from the user rest o f the Danish Dialogue System. (C) 1998 Academic Press Limited.