The scenario of the interaction between the user and the system is that the user has found a number of stories about an event that they want to follow in the future. The system uses these stories to "learn" what the story is about, and then alert the user when a new story is published about this event.

The graphic on the bottom depicts this scenario. The training data includes 4 stories known to be about and event and the test data are the future stories the the system searches through.

Tracking performance is effected by a number of factors: the number of training stories, the language of the training/test stories, the quality of the transcripts e.g. Automatic Speech Transcriptions vs. human generated transcripts. The evaluations seek to quantify these effects by provided adequate experimental controls.

To learn more about technology used to do TDT tracking, or to learn how well these systems perform, there were several system description papers submitted for the 1999 TDT workshop. Also, the TDT Research Links contains pointers to related tracking research.