The University of Queensland Homepage Link to my.UQUQ HomepageSearch UQFind all Contacts at UQStudy Options at UQNews at UQEvents at UQUQ MapsUQ Library
Go to advanced search page

Stochastic Methods for Optimization and Machine Learning

Botev, Zdravko I. (2005-11-01). Stochastic Methods for Optimization and Machine Learning Other, Department of Mathematics, School of Physical Sciences, The University of Queensland.

 
Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
Honors_Thesis_06N.ps   Honors_Thesis_06N.ps application/postscript 869.65KB 70
botev_thesis.pdf   botev_thesis.pdf application/pdf 570.32KB 432

Author(s) Botev, Zdravko I.
Thesis Title Stochastic Methods for Optimization and Machine Learning
School, Centre or Institute Department of Mathematics, School of Physical Sciences
Institution The University of Queensland
Publication date 2005-11-01
Thesis type Other
Subjects 230118 Optimisation
Abstract/Summary In this project a stochastic method for general purpose optimization and machine learning is described. The method is derived from basic information-theoretic principles and generalizes the popular Cross Entropy method. The effectiveness of the method as a tool for statistical modeling and Monte Carlo simulation is demonstrated with an application to the problems of density estimation and data modeling.
Keyword(s) maximum entropy
cross entropy
measures of information
Monte Carlo simulation
statistical modelling
machine learning
CE method
kernel smoothing
regularization theory
functional optimization
 
Versions
Version Filter Type
Access Statistics: 239 Abstract Views, 500 File Downloads Detailed Statistics
Created: Mon, 13 Feb 2006, 10:00:00 EST by Zdravko Botev on behalf of School of Information Technol and Elec Engineering. Detailed History