11_Chapter 6 Findings Conclusion and Implications of the study.pdf
Conclusion Key Findings
Transcript of Conclusion Key Findings
Abstract
Research Question
While cyber-attacks continue to linger, the Fourth Industrial Revolution's interest in Artificial Intelligence (AI), and the adaptation of Machine Learning (ML) and Deep Learning (DL) techniques into its framework, interpret cybersecurity issues soluble. Moreover, the dual computational problem-solving methods of Artificial Intelligence's Machine and Deep Learning structured framework placates Artificial Security Intelligence's defense viable against cybersecurity threats. This thesis explores the security feature optimization approaches of Ph.D. recipients and 2020 IntruDTree Model creators Sarker and Abushark's machine learning-based intrusion detection tree while producing promising results. Additionally, Sarker and Abushark optimized their security feature methodologies by minimizing complexity and achieving a variance analysis success rate of 98%, according to a two-time Ph.D. Honoree George W. Snedecor's chi-squared test. Therefore, by the IntruDTree Machine Learning-based model achieving a chi-squared statistical variance analysis success rate of 98%, cybersecurity solution, Artificial Security Intelligence, proves practical.
Methodology Exploratory Case Study: Investigate pragmatic solutions to cyber-attack phenomena using the limited
research of Machine Learning based Artificial Intelligence.
Why Machine Learning?The cyber vulnerability environment requires organizations to map and compare
millions of external and internal data points through their networks and users continuously (Cisco). It is just not possible to handle this amount of data for a small group of individuals (Cisco). Machine learning excels in this field as it can detect patterns and identify threats in large data sets at machine level (Cisco). Through automating the analysis, cyber teams can identify risks faster and distinguish cases that require more in-depth human analysis (Cisco).
Key Findings Fig. 1 An example of a Machine Learning based decision tree for detecting cyber
anomalies.
Fig. 2 Analysis of Variance (ANOVA) Equation
𝑟𝑟 𝑋𝑋,𝑌𝑌 =∑𝑖𝑖=1𝑛𝑛 (𝑋𝑋𝑖𝑖 − �𝑋𝑋)(𝑌𝑌𝑖𝑖 − �𝑌𝑌)
∑𝑖𝑖=1𝑛𝑛 (𝑋𝑋𝑖𝑖 − �𝑋𝑋)2 ∑𝑖𝑖=1𝑛𝑛 (𝑌𝑌𝑖𝑖 − �𝑌𝑌)2
Fig. 3 Chi-Squared Test
𝑋𝑋2 = �𝑖𝑖=1
𝑛𝑛(𝑂𝑂𝑖𝑖 − 𝐸𝐸𝑖𝑖)2
𝐸𝐸𝑖𝑖
Conclusion Artificial Security Intelligence research concentrated on:A.I. mechanisms and hypotheses linked to cyber security to facilitate synergy of
both.Build expertise and information focused on the convergence of A.I. and data
security.To investigate and facilitate intelligence modeling.
It is not a matter of if, but a matter of when A.I. becomes the frontrunner of cyber-attacks.The IntruDTree Machine Learning-based model will need to be an integral unit of
defense and in the framework of Artificial Security Intelligence.
Are there Artificial Security Intelligence solutions practical to solving cyber securityproblems?
Analysis The IntruDTree Machine Learning-based model achieved a chi-squared statistical
variance analysis success rate of 98% and proving Artificial Security Intelligence pragmatic.
I would like to acknowledge and thank Professor John Metzger for his continuedpatience and guidance throughout this Research Project. After realizing my initialresearch subject was not pragmatic, it was he who persisted I continue with a newsubject and made this project possible.
“Networking, Cloud, and Cybersecurity Solutions.” Cisco, 1 Apr. 2021,
www.cisco.com/c/en/us/index.html.
Sarker, Iqbal H., et al. “IntruDTree: A Machine Learning-Based Cyber Security
Intrusion Detection Model.” 2020, doi:10.20944/preprints202004.0481.v1.
Acknowledgements
References