Detection of signals by Monte Carlo singular spectrum analysis: Multiple testing

Abstract

Detection of a signal in a noisy time series using Monte Carlo singular spectrum analysis (MC-SSA) is studied from the statistical viewpoint. The MC-SSA test consists of simultaneous testing of several hypotheses related to the presence of different frequencies. The multiple MC-SSA test procedure is constructed to control the family-wise error rate. The technique to control both the type I and the type II errors and also to compare criteria is proposed to study several versions of MC-SSA.