Stochastic Surfaces in the Least Squares Wavelet Analysis

Ebrahim Ghaderpour, York University, Toronto, ON, Canada and Spiros D Pagiatakis, York University, Department of Earth and Space Science, Toronto, ON, Canada

Contact First Author: Ebrahim Ghaderpour; ebig2@yorku.ca

Previously Published Material: This paper is under review by Digital Signal Processing (Elsevier).

Abstract ID#: 34058

 

English Abstract:
The stochastic significance of peaks of the least-squares spectrum was discussed by Pagiatakis (1999), which we extend to define the stochastic surfaces of the Least-Squares Wavelet Analysis (LSWA) spectrograms that follow the beta distribution. Several examples will be presented to show the stochastic surfaces for the LSWA spectrograms that define confidence levels (usually 95% or 99%) above which spectral peaks in LSWA spectrograms are statistically significant. In the examples, we also show that the spectral peaks must be stronger for higher frequencies to be statistically significant as the number of data points for the segments of a time series decreases in the LSWA.
Back to: Big Data and Advanced Geocomputations