Application of Pseudospectrum Estimation Method Based on Singular Vectors for the Analyzing Cyclic Behavior of the Temperature in an Office Building

Abstract

Spectral analysis of time series has a huge application in different fields of science, starting from the Radar technologies to vibration processes, economics and finance. Recent research proved that left and right singular vectors of nonstationary time series consist of periodic deterministic components and white noise, have an equivalent pseudospectral
structure of original time series. In the given article, application of this property has been demonstrated on the base of analysis of cyclic behavior of the temperature in an office building.