ClusterID:ImportantPointsToRemember

This ClusterID package has many levels and sometimes the novice can lose track of what is going on and make a mistake. On this page is just a list of things to look over if you are having a problem. Maybe something here will pop out. To make this list useful, users must send us useful items to add to this list. As the developers we don't know what is possible to miss and cause confusion.

1) Remember that a net is trained for a particular detector configuration. Running a net for the SD detector on an LD detector will not give good results. However, running an SD net on a detector which is very similar to SD may give reasonably good results.

2)Remeber that a net is trained with a specific data set. A good study for someone to do is how important this is. For example, if the net was trained on particles representing distributions and energies at the Zpole and it is used on ZZ events at 1 TeV how well will it work? As a first approximation it should be fine but as we require more accurate resolution this may become an issue.