Event and Analysis Classes
Analysis Classes
We have six analysis classes:
- classA (also known as GoodEvent3)
- all (also known as GoodEvent1)
- FrontA
- BackA
- FrontB
- BackB
Note: The two classes most suitable for science analysis are:
- Class A is the highest quality, with:
- Best PSF
- Lowest background contamination.
- all
- More effective area, but
- more background
These can be selected in any of the Science Tools which prompt for an IRF as DC2_A
(for class A) and DC2 (for all).
Notes:
- Instrument Response Functions. IRFs have been constructed so that each event belongs to one - and only one - event class. Event classes are combined to make the analysis classes that you select for analysis.
- ClassB Events. ClassB events are events which belong to All, but are not classA events.
- Front or Back. Each event class is further subdivided into front and back (depending on whether the event converted in the front or back section of the LAT).
- Individual Event Classes. Each event class can also be studied individually by selecting:
- DC2_FA (where FA = Front Event Class A)
- DC2_BA (where BA = Back Event Class A)
- DC2_FB (where FB = Front Event Class B)
- DC2_BB (where BB = Back Event Class B)
Detailed cuts developed from the background analysis and PSF optimization are described in the Confluence wiki page: Trigger and Filter Setting and Event Classes.
These are stored in the FTI in the EVENT_CLASS column as an integer ranging from 0-4.
- [0-1] – A Front, Back
- [2-3] – B Front, Back
The EVENT_CLASS column in FT1 is defined using a set of TCuts on the merit summary ntuple.
Note: Three EVENT_CLASS variables not defined in current FT1 documentation, but whose definitions are obvious, should also be mentioned here; these are:
EVENT_Class
RUN_ID
EVENT_ID
Choices that make sense for use with the Instrument Response Functions (IRF’s) are:
- All
- Any individual value
- 0 AND 1 (i.e., A only)
Also See
Last updated by: Chuck Patterson
02/28/2006 |
|
|