Dissertation/Thesis Abstract

Bayesian Networks with an application to credit risk
by Kent, Geoffrey, M.S., California State University, Long Beach, 2014, 173; 1526920
Abstract (Summary)

Bayesian Networks are a local factorization of the joint probability distribution, combined with a directed acyclic graph that represents conditional independence statements. Bayesian networks have become a pervasive methodology in the modeling of uncertain domains because Bayesian networks systematically describe and simplify complex dependencies between random variables in an efficient and principled manner. Founded on the centuries-old Bayesian probability theory, advances in statistics and computer science have allowed Bayesian networks to be built and applied over a wide range of applications in an efficient and intuitive manner. Many examples are presented to demonstrate the capabilities of the networks, including applications to handwriting analysis, medical diagnosis, recommender systems, ecommerce, and credit risk.

Indexing (document details)
Advisor: Watson, Saleem
Commitee: Safer, Alan, Suaray, Kagba
School: California State University, Long Beach
Department: Natural Sciences and Mathematics
School Location: United States -- California
Source: MAI 53/06M(E), Masters Abstracts International
Source Type: DISSERTATION
Subjects: Statistics, Artificial intelligence
Keywords: Bayesian Networks, Credit risk, Handwriting analysis, Probablistic graphical models
Publication Number: 1526920
ISBN: 9781321277203
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest