Using a Multi-Variate Statistical Classifier in a Gravitational Wave Search

The principal problem of gravitational wave detection is distinguishing true gravitational wave (GW) signals from non-Gaussian noise artifacts. Past searches for GWs developed \emph{ad hoc} detection statistics in an attempt to separate the expected GW signals from detector glitches.  The high dimension parameter spaces of GW waveforms pose a problem to the human analyst when trying to develop optimal detection statistics. Multi-Variate Statistical Classifiers have been used in computer science and other branches of physics to directly probe high dimensional spaces and distinguish between classes of events.  We will describe the Random Forest of Bagged Decision Trees algorithm and demonstrate its efficacy at distinguishing simulated GW signals from detector noise in the context of a search for GWs from black hole ringdown in LIGO data.