Study of statistical properties of hybrid statistic in coherent multi-detector compact binary coalescences Search

Abstract: In this article, we revisit the problem of coherent multi-detector search of gravitational wave from compact binary coalescence with Neutron stars and Black Holes using advanced interferometers like LIGO-Virgo. Based on the loss of optimal multi-detector signal-to-noise ratio (SNR), we construct a hybrid statistic as a best of maximum-likelihood-ratio(MLR) statistic tuned for face-on and face-off binaries. The statistical properties of the hybrid statistic is studied. The performance of this hybrid statistic is compared with that of the coherent MLR statistic for generic inclination angles. Owing to the single synthetic data stream, the hybrid statistic gives low false alarms compared to the multi-detector MLR statistic and small fractional loss in the optimum SNR for a large range of binary inclinations. We have demonstrated that for a LIGO-Virgo network and binary inclination, \epsilon < 70 deg. and \epsilon > 110 deg., the hybrid statistic captures more than 98% of network optimum matched filter SNR with low false alarm rate. The Monte-Carlo exercise with two distributions of incoming inclination angles namely, U[cos(\epsilon)] and more realistic distribution proposed by B. F. Schutz are performed with hybrid statistic and gave ~5% and ~7% higher detection probability respectively compared to the two stream multi-detector MLR statistic for a fixed false alarm probability of 10^-5.