Reusable dialog component for content selection from large data sets

Inherent limitations of spoken language interfaces make the task of information access from large data sets difficult. Providing a dialog component which can be easily configured to access information from such data sets is immensely useful. Such component would ease and expedite the development of speech applications. We propose a dialog component which makes use of user preferences, user profile and utterance history to select relevant information from large data sets. Content presentation is also determined by user preferences and utterance history. The evaluation shows the effectiveness of the technique and effect of user profile in accessing information. It also demonstrates reusability of component to access different datasets