Methods to characterize non-Gaussian noise in TAMA

We present a data characterization method for the main output signal of the interferometric gravitational-wave detector, in particular targetting at effective detection of burst gravitational waves from stellar core collapse. The time scale of nonGaussian events is evaluated in this method, and events with longer time scale than real signals are rejected as non-Gaussian noises. As a result of data analysis using 1000 h of real data with the interferometric gravitational-wave detector TAMA300, the false-alarm rate was improved 103 times with this non-Gaussian noise evaluation and rejection method.