Multiple Rebin Quest

In addition to the cross-check with other GRB catalogs, another advantage concerning the off-line quest with respect to the on-line case, is offered by the multiple rebin quests. While in the on-line case, the search algorithms apply only to the original GRBM 1 s light curves (for the motivations already described), in the off-line case the same algorithms, albeit with different values of the parameters, also applied to the rebinned light curves: 1, 2, 4 and 8 s light curves have been scanned. Nevertheless, the moving interval used for background fitting, keeps the original 1 s step; in other words, when, for example, the scan of the 8 s light curves is performed, each original 1 s bin contributes to 8 different 8 s bins, that are scanned apart, till the 8 s scanned bin has passed on. This particular choice is aimed to increase the number of potential off-line triggers, that can be discovered. The parameters defining the intervals used for background fitting (eq. ) are reported in table ; concerning the detection thresholds, these have been left unchanged from the 1 s off-line case (eqq. , , ).

Table: Quest Parameters adopted for the different rebin times

Time Bin	$N_b$	$N_a$	$n_b$	$n_a$
(s)
1	60	30	65	91
2	50	60	60	41
4	25	25	30	25
8	15	20	20	20

The effect of a multiple rebin quest is given by an increase of the number of off-line triggers, and this property improves the algorithms sensitivity; on the other hand, by increasing the number of off-line triggers, the amount of human work, necessary to control them one by one by visual inspection, increases too. Nevertheless, one limit of the multiple rebin quest is that the number of false off-line triggers that are automatically found, is proportionally greater than in the case of the only 1 s quest; in other words: on the whole, the multiple rebin quest has a lower reliability than the 1 s quest: for example, at 8 s, one of the largest classes of transient events, that mimic the occurrence of a burst, is given by the occultation steps of strong hard X-ray sources, like the Crab Nebula and Cygnus X-1 (section ). About this, an interesting method has been applied to the BATSE continuous data ([Stern et al., 2000b]) for automatically discarding the occultation steps due to Cygnus X-1, and, in the near future, it could be adjusted for the GRBM data scan for more sensitive searches.

The results of these quests are reported and discussed in the next chapter, entirely devoted to the catalog of GRBs extracted from the GRBM data, and that have been caught by means of the quests here described.