A NEW REPRESENTATION OF TRENDLESS SEQUENCES: TESTED ON SEISMOLOGICAL DATA

Abstract

This paper proposes a new method for representing detrended random sequences (DRS) based on an analysis of the distribution of the number of level crossings L (or the number of zeros
of the function y(t) − L with a sign change), where the level L sequentially changes between the minimum and maximum values of a given detrended sequence. As it turns out, upon closer analysis, this representation transforms DRS into a pseudo-resonance curve, which can be described by 3-10 parameters, depending on the depth of DRS analysis. This transformation allows for comparison of different DRSs without relying on model representations of the distribution of random points within the DRS.

In seismology, the problem of distinguishing recorded earthquake seismograms from explosions of
artificial or man-made origin is central. The method proposed in this paper, based on explicit physical qualitative features and using more detailed correlation methods, allows for a reduction and qualitative improvement in the set of features used to distinguish earthquake seismograms from man-made explosions (MEs). Validation against an accessible database enables this distinction to be made and recommends this method for widespread use.