Integral and Series Representation of Infinitely Divisible Processes with Applications to Their Prediction and to Their Sample Path, Statistical and Structural Properties
Abstract
The Gaussian model has been used to describe many random phenomena in science and engineering because of its versatility and mathematical simplicity. However, the Gaussian model is not universally applicable; and, in fact, there are many instances, both in the areas of theoretical research and engineering applications, where the need of non-Gaussian models, particularly those with infinite variance, can be identified. For instance, man-made noise in a hostile environment can be made to depart from Gaussian behavior; and the natural noise occurring in situations where weak signals need to be extracted also tends to not follow the Gaussian pattern. These are but a few examples that might be of particular interest to defense agencies, where non-Gaussian modeling and their analysis are most desirable. Keywords: Gaussian models; Integral and series representation; Prediction; Sample path; Statistical and structural properties; Man-made noise; Statistical processes; White noise; Acoustics.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 31, 1989
- Accession Number
- ADA219625
Entities
People
- Balram S. Rajput
- Jan Rosinski
Organizations
- University of Tennessee