So that we can score pipelines based on how it does relative to that ground truth information. This is probably complicated, as the form of ground truth data may change based on the pipeline / classification type as well as the post-processing done. For instance, we may want to have a way to mark ground truth as regions / intervals and score pipelines against the percentage of the intervals that match. But we might also want a way to specify ground truth as events and score the pipeline by distance between its events and the ground truth ones.
So that we can score pipelines based on how it does relative to that ground truth information. This is probably complicated, as the form of ground truth data may change based on the pipeline / classification type as well as the post-processing done. For instance, we may want to have a way to mark ground truth as regions / intervals and score pipelines against the percentage of the intervals that match. But we might also want a way to specify ground truth as events and score the pipeline by distance between its events and the ground truth ones.