Title: Results of the 19972000 Search for Burst Gw by IGEC
1Results of the 1997-2000 Search for Burst Gw by
IGEC
G.A.Prodi - INFN and Università di Trento,
Italy International Gravitational Event
Collaboration
http//igec.lnl.infn.it
GWDAW 2002
ALLEGRO group ALLEGRO (LSU) http//gravity.phys.
lsu.edu Louisiana State University, Baton Rouge
- Louisiana
AURIGA group AURIGA (INFN-LNL) http//www.auriga
.lnl.infn.it INFN of Padova, Trento, Ferrara,
Firenze, LNL Universities of Padova, Trento,
Ferrara, Firenze IFN- CNR, Trento Italia
NIOBE group NIOBE (UWA) http//www.gravity.
pd.uwa.edu.au University of Western Australia,
Perth, Australia
ROG group EXPLORER (CERN)
http//www.roma1.infn.it/rog/rogmain.html NAUTIL
US (INFN-LNF) INFN of Roma and LNF Universities
of Roma, LAquila CNR IFSI and IESS, Roma -
Italia
2OUTLINE
overview of the EXCHANGED DATA SET
1997-2000 sensitivity and observation
time candidate burst gw events
multiple detector DATA ANALYSIS directio
nal search strategy search as a function of
amplitude threshold false dismissal or
detection efficiency estimation of accidental
coincidences by time shifts
methods ? L.Baggio tomorrow
RESULTS accidental coincidences are
Poisson r.v. compatibility with null
hypothesis upper limit on the rate of detected
gw unfolding the sources (not yet)
3DETECTOR LOCATIONS
almost parallel detectors
4EXCHANGED PERIODS of OBSERVATION 1997-2000
ALLEGRO
AURIGA
NAUTILUS
EXPLORER
NIOBE
fraction of time in monthly bins
Fourier amplitude of burst gw
arrival time
5DIRECTIONAL SEARCH
amplitude (Hz-1)
time (hours)
amplitude (Hz-1)
time (hours)
6DATA SELECTION
amplitude (Hz-1)
time (hours)
7OBSERVATION TIME 1997-2000
total time when exchange threshold has been lower
than gw amplitude
8DATA SELECTION
amplitude (Hz-1)
time (hours)
amplitude (Hz-1)
time (hours)
9RESULTING PERIODS of OBSERVATION and EVENTS
no directional search
time (hours)
directional search
time (hours)
10AMPLITUDE DISTRIBUTIONS of EXCHANGED EVENTS
normalized to each detector threshold for trigger
search
typical SNR of trigger search thresholds
? 3 ALLEGRO, NIOBE ? 5 AURIGA, EXPLORER,
NAUTILUS amplitude range much wider than
expected non modeled outliers dominating at high
SNR
11FALSE ALARM REDUCTION
by thresholding events
natural consequence AMPLITUDE CONSISTENCY of
SELECTED EVENTS
12FALSE DISMISSAL PROBABILITY
- data selection as a function of the common
search threshold Ht - keep the observation time when false dismissal
is under control - keep events above threshold
- efficiency of detection depends on signal
amplitude, direction, polarization - e.g. gt 50 with amplitude gt Ht at each detector
- time coincidence search
- time window is set requiring a conservative
false dismissal - robust and general method Tchebyscheff
inequality
false alarms ? k
- amplitude consistency check gw generates events
with correlated amplitudes - testing (same as above)
? efficiency of detection versus false
alarms maximize the ratio
best balance in our case time coincidence max
false dismissal 5 ? 30 no rejection based
on amplitude consistency test
13POISSON STATISTICS of ACCIDENTAL COINCIDENCES
Poisson fits of accidental concidences ?2 test
sample of EX-NA background one-tail probability
0.71
agreement with uniform distribution
histogram of one-tail ?2 probabilities for ALL
two-fold observations
14SETTING CONFIDENCE INTERVALS
- unified frequentistic approach
- tomorrow talk by L. Baggio
- References
- B. Roe and M. Woodroofe, PRD 63, 013009 (2000)
- most likely confidence intervals ensuring a
given coverage (our choice) - 2. G.J.Feldman and R.D.Cousins, PRD 57, 3873
(1998) - Recommendations of the Particle Data Group
http//pdg.lbl.gov/2002/statrpp.pdf - see also the review F.Porter, Nucl. Instr. Meth
A 368 (1996) - COVERAGE probability that the confidence
interval contains the true value -
- unified treatment of UPPER LIMIT ?? DETECTION
- freedom to chose the confidence of goodness of
the fit tests independently from the confidence
of the interval
15SETTING CONFIDENCE INTERVALS / 2
GOAL estimate the number of gw which are
detected with amplitude ? Ht
Example confidence interval with coverage ? 95
16SETTING CONFIDENCE INTERVALS / 3
systematic search on thresholds many trials
! all upper limits but one
- testing the null hypothesis
- overall false alarm probability 33
- at least one detection in case
- NO GW are in the data
PDG recommendation A potential difficulty with
unified intervals arises if, for example, one
constructs such an interval for a Poisson
parameter s of some yet to be discovered signal
process with, say, 1 - ? 09. If the true
signal parameter is zero, or in any case much
less than the expected background, one will
usually obtain a one-sided upper limit on s. In a
certain fraction of the experiments, however, a
two-sided interval for s will result. Since,
however, one typically chooses 1 - ? to be only
09 or 095 when searching for a new effect, the
value s 0 may be excluded from the interval
before the existence of the effect is well
established. It must then be communicated
carefully that in excluding s 0 from the
interval, one is not necessarily claiming to have
discovered the effect.
NULL HYPOTHESIS WELL IN AGREEMENT WITH THE
OBSERVATIONS
17UPPER LIMIT /1
on RATE of BURST GW from the GALACTIC CENTER
DIRECTION with measured amplitude ? search
threshold no model is assumed for the sources,
apart from being a random time series
rate year -1
ensured minimum coverage
search threshold Hz -1
true rate value is under the curves with a
probability coverage
18UPPER LIMIT /2
on RATE of BURST GW without performing a
directional search measured amplitude ? search
threshold (amplitudes of gw are referred to the
direction of detectors) no model is assumed for
the sources, apart from being a random time series
rate year -1
ensured minimum coverage
search threshold Hz -1
true rate value is under the curves with a
probability coverage