COMING SOON! PQDT Open is getting a new home!

ProQuest Open Access Dissertations & Theses will remain freely available as part of a new and enhanced search experience at

Questions? Please refer to this FAQ.

Dissertation/Thesis Abstract

Applicability of Predictive Modeling on a Network Enterprise Risk Management Critical Infrastructure
by Valentin Torres, Ana Hilda, D.Sc., Marymount University, 2020, 146; 28391145
Abstract (Summary)

Bayesian Networks can identify, correlate, and predict causal relationships among network risk factors and improve network performance The application of Bayesian Networks supports risk evaluation, research and development of new products, and Bayesian solutions for enterprises predicting the strategic marketing and management decisions as required by the National Institute of Standards and Technology Interagency Report — 8286 released in October 2020.

The purpose of the research is to examine whether correlational and predictive modeling techniques can identify, correlate, and predict network performance from casual events among current status, customer impact, network failure, network impact, outage, and source of impact.

The population sample of this exploratory quantitative research consists of 410 network risk incident events, including six network risk incidents for a total of 2,460. The data analyses do not apply to missing data correction factors. The research applies a simple random sampling probability of selection, assigning an equal probability of selection to each network risk incident event.

This research applies two different statistical analyses: correlation analyses and predictive modeling techniques. The correlation analyses consist of the Fault Tree Analysis, Generalized Linear Model for Poisson, and Generalized Linear Model for Gamma-Poisson.; meanwhile, the predictive modeling techniques consist of Multiple Logistic Regression, Bayesian Generalized Linear Model, and Bayesian-Monte Carlo Generalized Linear Model. The research methodology applies correlation analyses to determine whether two or more network risks are strongly associated with establishing a causal outcome. Once a causal relationship among two or more network risks throughout correlational analyses, organizations may apply predictive modeling techniques to predict the likelihood of an expected outcome of network risk incidents in the function of one or other network risk incident criteria. The results of the research did not show significant evidence to reject any of the null hypotheses from Hi to H12 with a confidence interval of 95% and a p-value less than 0.05.

In conclusion, further research shall delineate a quantitative approach for enterprise risk management compliance with the National Institute of Standards and Technology Interagency Report - 8286. This research did not consider any financial or monetary data to quantify the network performance's economic impact provoked by network risk incidents. The financial data limited the capacity to forecast the impact likelihood of a network risk incident in the network performance. Keywords: predictive modeling techniques, enterprise network risk management, NISTIR-8286, network enterprise network risk management

Indexing (document details)
Advisor: Schaeffer, Donna M.
Commitee: Mbaziira, Alex V., Munoz, Jose L.
School: Marymount University
Department: Cybersecurity
School Location: United States -- Virginia, US
Source: DAI-B 82/8(E), Dissertation Abstracts International
Subjects: Computer science, Information Technology
Keywords: Enterprise Network Risk Management, NISTIR-8286, Network Enterprise Network Risk Management
Publication Number: 28391145
ISBN: 9798569969906
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy