Presented is a new algorithm for implementing the Relative Entropy method for estimating the probability density function of a process. The Relative Entropy principle can also be used to estimate the power spectrum of a random process as presented in Appendix A.
Chapter 2 presents the current method of pdf estimation using the Relative Entropy method. In this method a set of Lagrange multipliers are solved for simultaneously. The new algorithm, discussed in Chapter 3, presents a new, efficient method for solving for one Lagrange multiplier at a time, iteratively. The procedure presented allows one to estimate the probability density function of a process based on a prior estimate of the pdf of the signal.
The performance of the method is tested using known probability density functions. Also, the performance of the new method is compared with that of the Maximum Entropy based method. The estimation of a pdf of a speech signal will also be used to demonstrate the performance of the Relative Entropy method using the new algorithm.
If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at STARS@ucf.edu
Alsaka, Yacoub A.
Master of Science (M.S.)
College of Engineering
Length of Campus-only Access
Masters Thesis (Open Access)
Marinelli, William A., "A New Relative Entropy Based Method for Probability Mass Function Estimation" (1987). Retrospective Theses and Dissertations. 5075.
Contributor (Linked data)