1Chapter One
Introduction
Overview
Cyber threats have been in existence in the modern world. Machine learning algorithms
have since been adopted to handle cyber threats and improve cyber security in organizations.
cyber security is a combination of different technologies designed to protect computers and
systems. The use of machine learning algorithms tends to remain ambiguous concerns have been
raised about the use of algorithms to handle cyber threats. However, the exact operation of
machine learning algorithms by an organization is uncertain, and some critics have raised
concerns about the use of algorithms to neutralize cyber threats. Yet machine learning algorithms
have already overcome these threats effectively. In fact, logistic regression learning algorithm
has multiple applications: identifying threats, improving antivirus software, and fighting
cybercrime with artificial intelligence capabilities (Aravindan et al., 2020).
Introduction
Background and Problem Statement
Recent research has seen attempts to incorporate machine learning within cyber security,
but success has been gradual. With the advent of big data, machine learning has become more
promising in recent years. Organizations use machine learning algorithms for cyber security to
ensure their systems are safe from Distributed Denial-of-service (DDoS) attacks and other types
of cyber-attacks. The algorithms effectively detect and respond to attacks, improving the security
infrastructure of systems. On the other hand, machine learning algorithms can be weak, and, in
recent studies, researchers have found that such algorithms have been targeted by malicious
actors (Zhang et al., 2020).
2
According to Zhang et al. (2020), an organization must focus on combining artificial
intelligence and machine learning to ensure any vulnerabilities in the system are handled
effectively. Machine learning for cybersecurity is necessary for preventing cyber-attacks and
responding to changing behavior in real time.
Cyber-attacks have been responsible for the loss of essential data in many organizations
because there were no effective ways of combating the attacks. According to (Bilen & Ozer,
2021), organizations have focused on using machine learning algorithms, but the effectiveness of
machine learning in bolstering cyber security is still unknown. The lack of an understanding of
the effectiveness of machine learning contributes to unsecure systems in an organization and the
increased severity of cyber-attacks. The result of this is the slow implementation of cyber
security measures, ultimately undermining the detection and prevention of cyber-attacks.
Problem Statement
You need to add to your problem statement approximately 250 to 300 words)
Note: Articulation of a concise problem statement is key to a successful
proposal/dissertation manuscript. The problem statement is a brief discussion
of a problem or observation succinctly identifying and documenting the need
for and importance of the study.
. Present general issue/observation that in theory or practice leads to
the need for the study (in most cases citations within the last 5 years should
be included
Present a focused problem that leads to the need for a research
response. For some degree programs (DBA, EdD) the problem identified might
be a practical problem or issue in an organization.
Purpose of the Study
The purpose of this quantitative study is to evaluate the effectiveness of logistic
regression machine learning algorithm approaches for cyber security purposes. The study also
3
aims to understand how organizations improve cyber security through employing the best
possible cyber-attack defense (Al-Mhiqani et al., 2020).
What you need to DO:• Research method is identified as qualitative, quantitative, or
mixed method
• Research design is clearly stated and is aligned with the
problem statement.
• Specific population of proposed study is identified. The
number of participants that will serve as the sample should
be estimated based on a power analysis (quantitative/mixed
method) or conventions (qualitative
Significance of the Study
This research is significant because it provides insights into the significance of machine
learning in handling cyber threats. (Rashid et al., 2020) argue that for machine learning models
to be effective, they must learn to do various activities. From the statistics presented in this
study, there is a determination of how effective machine learning is compared to other cyberattacks countermeasures. Organizations can use the findings to determine how cyber security can
be improved by adopting machine learning. Throughout the study, there is an examination of the
cyber-attacks and the countermeasures adopted to handle the attacks. Specifically, the emphasis
is on the effectiveness of logistic regression machine learning algorithm approaches.
This study’s results will benefit different fields affected by cyber threats. As a result of
the increased use of the internet, there have been a high number of cyber-related attacks.
Concerning this, the findings of this study can be applied to eliminate cyber threats by using
4
machine learning algorithms. Machine learning threats are the bias, security, safety, and
inequality concerns related to machine learning technology. In the context of an organization,
machine learning threatens the available systems by making them porous. Cyber attackers have
capitalized on the complexity of machine learning technology to target different organizational
systems.
Machine learning threats could also threaten the organization’s systems due to the
increased possibility of error. In centralized and decentralized organizations, machine learning
algorithms will significantly impact the decision-making process in the future (Vergne, 2020).
The decision made to the machine learning algorithms depends on the effectiveness of the
algorithms in handling cyber threats. To this, the future of digital platforms depends on the
effectiveness of machine learning algorithms. In addition, the future of centralized and
decentralized organizations’ security largely depends on the success of machine learning
algorithms.
Considering the logistic regression machine learning algorithm approaches, the study’s
findings are significant in deciding whether to adopt the machine learning algorithm approaches
to improve cyber security in an organization. Also, the scope of the algorithm approaches
influences the decision made on whether to use machine learning for short or long-term
purposes. The study findings are significant in improving the current cyber security approaches
in an organization. Based on the data presented, organizations can determine the specific cyber
security areas that can be improved due to the use of machine learning approaches. With this, the
current practices adopted in handling DDoS attacks can be changed. Machine learning is vital in
ensuring an increased stock of knowledge in an organization (Sturm et al., 2021). Based on this,
the application of research carried out on machine learning is vital not only to the personnel but
5
also to the organization at large. This study can also be applied to coordinate machine learning
and organizational productivity. Productivity is majorly on the security and achievement of set
goals.
Two research questions guided this research: Are machine learning algorithms concerned
with detection of cyber threats? Is there a relationship between the application of machine
learning techniques and the mitigation of cyber threats? Exploring these research questions
provides the scope of machine learning algorithms in an organizational setting.
What you need to DO:Discussion demonstrates why the study is important and describes
the contribution(s) that the completion of the research makes to
the field of study.
Research Questions
RQ1. Are machine learning algorithms concerned with detection of cyber threats?
RQ2. Is there a relationship between the applications of machine learning techniques?
and the mitigation of cyber threats?
The hypothesis:
The null hypothesis (H0) is True: There exists no meaningful relationship between
machine learning algorithms and the detection of cyber threats.
The alternative hypothesis (H1) is False: There exists a significant relationship between
machine learning algorithms and the detection of cyber threats.
What you need to Do:-
6
Theoretical Framework
The initial stage in data collection was to visit the various stations of the participants,
namely, their places of employment. The second step was to seek approval from both the
organization and the study participants. Consent is crucial for allowing the data gathering process
to continue (Jilcha Sileyew, 2020). The third step was to choose a specific time when the
participants would be accessible for data collection. The researcher collected the participants’
data in the final stage after presenting multiple survey questions to the participants online. In this
study, the author used surveys, interviews, and controlled observations to collect data.
The author chose a total of 82 IT professionals as the participants. To identify the
correlations between the category variables of various types, the researcher employed the
ANOVA statistical test. The ANOVA returns the P-value and f-test score. The f-test calculates
the variance between group means within each group mean (Faulkenberry & Wagenmakers,
2020).
According to (Guo et al., 2021), machine learning technology has gained interest in
modern times and is being applied for purposes that were unimaginable in the previous years.
Different applications of machine learning technology have contributed to increased privacy
protection. Based on this, the handling of cyber threats has been improved by the use of machine
learning. Research tends to build on the significance of machine learning by focusing on
improving security. On this, machine learning algorithms have proved significant in developing
new solutions to mitigate cyber-attacks.
Although there is an increase in the use of machine learning algorithms, uncertainty about
the effectiveness of the algorithms tends to exist. To determine how effective the machine
learning algorithms are, the focus is on how effective they are in intrusion detection. As brought
7
out by (De Carvalho Bertoli et al., 2021), modern intrusion detection solutions are complex and
involve a combination of different features. Therefore, organizations have invested in adopting
technologies to simplify intrusion detection.
Machine learning algorithms have been determined to be effective in intrusion detection.
However, machine learning algorithms are only effective in specific organizations. The success
of machine learning algorithms in specific organizations has contributed to increased certainty of
machine learning algorithms. (Guo et al., 2021) argues that machine learning algorithms have
gained renewed interest due to the availability of the technology powering the algorithms. From
a cybersecurity angle, machine learning algorithms have brought about specific opportunities and
challenges for different applications.
The effectiveness of machine learning algorithms is evident in intrusion detection in IoT
devices. On the Internet of Things (IoT) field, the rising concerns are anomaly detection and
increased threats in IoT infrastructure. (Morfino & Rampone, 2020) argue that machine learning
algorithms are effective in the identification of IoT systems attacks. The effectiveness of the
algorithms is determined by focusing on the application performance. Concerning this, machine
learning algorithms have a high accuracy in detecting intrusion in IoT systems.
Machine learning algorithms are ideal for the detection of artificial intelligence-based
malware. Machine learning algorithms have been determined to be productive solutions for
artificial intelligence-based malware. Therefore, the algorithms have been used to eliminate
problems that relate to cyber threats. The effectiveness of the algorithms for cyber threats has
been contributed by the significant advancements made in the technology. Although machine
learning algorithms have improved in the detection of cyber threats, technology advancement has
8
played a part in allowing attackers to work around the available intrusion detection systems (Patil
et al., 2021)
In malware detection, there is uncertainty in the productivity of machine learning.
Therefore, organizations have invested in different models of malware detection. (Shaukat et al.,
2020) notes that there are challenges that arise as a result of the use of machine learning
algorithms in organizations. The majority of the challenges that arise with the use of machine
learning algorithms have been contributed by the increased dependence on the internet. In
addition to internet dependence, the complexity of machine learning algorithms has proved to be
a significant challenge.
What you need to DO:The Theoretical Framework you have given, you need to update it
•
You need to add the broad conceptual and/or theoretical area under
which the research falls and how proposed research would fit
within other research in the field. Discussion specifically includes
important issues, perspectives, and, if appropriate, controversies
in the field
•
Discussion reflects knowledge and familiarity with both historical
and current literature.
Limitations of the Study
This study is characterized by a number of limitations. This study is limited in relevance
to only professionals in the information technology departments and not the general population.
Among the limitations are insufficient data on organizations’ use of machine learning models.
The insufficient data is contributed by the slow adoption of the technology by organizations and
the limited number of professionals using machine learning algorithms. Another limitation is
9
time constraints. Time constraints undermined the in-depth analysis of the research problem as
most participants had limited time to provide important information on machine learning.
Additionally, this research has only studied organizations that received some form of cyberattack. The researcher was also only interested in the safest approach to fend off cyber threats
(Estévez-Pereira et al., 2020). Biased views formed part of the limitations of the study. A number
of the participants were biased in that they selected the data that supported their main arguments.
Limited access to information was also a limitation. There was limited access due to the
complexity of accessing specific organizations. This contributed to making significant changes in
the research design.
Assumptions
The study was constructed with the following assumptions:
1. Cyber-attacks and network intrusion were significant detriments to the performance of
organizations. The assumption is based on the findings from Aravindan et al. (2020)
2. The participants have the necessary cyber security knowledge based on their
qualifications and experience.
3. The participants provided accurate and honest responses to the questions presented.
4. All the participants have experienced DDoS attacks in different settings.
5. The selected organizations have had episodes of cyber-attacks prompting the adoption of
countermeasures.
Definitions
Cyber-attack. A hacker’s attempt to destroy or damage a computer or the system of an
organization (Aravindan et al., 2020).
10
Cyber security. The practice of protecting servers, electronic systems, mobile devices,
and networks from malicious attacks (Aravindan et al., 2020).
Network intrusion. The existence of an unauthorized activity on a specific network
(Zhang et al., 2020).
Data science. A field that uses scientific processes, methods, and algorithms to extract
necessary knowledge from unstructured and structured data (Sarker et al., 2020).
Summary
With innovations in intelligent systems, machine learning has become a critical tool in
eradicating cyber threats. The researcher chose to study machine learning algorithms to
determine the most effective method of combating a real problem for many organizations. This
chapter was divided into ten sections. The introduction provided information on machine
learning and cyber security. Then the background information on machine learning discussed
how machine learning algorithms are becoming a popular method to combat cyber threats, but it
is still uncertain how effective these algorithms are. The next section provided an analysis of the
population and the population sample. The methodology section described the quantitative
methods used in this study. The alignment section gave an overview of the research questions,
problem statement, statement of purpose, and the study’s title. The next chapter will review the
literature on this topic.
11
References
Aravindan, C., Frederick, T., Hemamalini, V., & Cathirine, M. V. J. (2020). An Extensive
Research on Cyber Threats using Learning Algorithm. 2020 International Conference on
Emerging Trends in Information Technology and Engineering (ic-ETITE), 1–8.
https://doi.org/10.1109/ic-ETITE47903.2020.337
Al-Mhiqani, M. N., Ahmad, R., Zainal Abidin, Z., Yassin, W., Hassan, A., Abdulkareem, K. H.,
Ali, N. S., & Yunos, Z. (2020). A Review of Insider Threat Detection: Classification,
Machine Learning Techniques, Datasets, Open Challenges, and Recommendations.
Applied Sciences, 10(15), 5208–. https://doi.org/10.3390/app10155208
Bilen, A., & Özer, A. B. (2021). Cyber-attack method and perpetrator prediction using machine
learning algorithms. PeerJ. Computer Science, 7, e475–e475.
https://doi.org/10.7717/peerj-cs.475
De Carvalho Bertoli, G., Pereira Junior, L. A., Saotome, O., Dos Santos, A. L., Verri, F. A.,
Marcondes, C. A., Barbieri, S., Rodrigues, M. S., & Parente De Oliveira, J. M. (2021).
An end-to-end framework for machine learning-based network Intrusion Detection
System. IEEE Access, 9, 106790–106805. https://doi.org/10.1109/access.2021.3101188
Estévez-Pereira, J. J., Fernández, D., & Novoa, F. J. (2020). Network Anomaly Detection Using
Machine Learning Techniques. Proceedings, 54(1), 8–.
https://doi.org/10.3390/proceedings2020054008
Faulkenberry, Ly, A., & Wagenmakers, E.-J. (2020). Bayesian Inference in Numerical
Cognition: A Tutorial Using JASP. Journal of Numerical Cognition, 6(2), 231–259.
https://doi.org/10.5964/jnc.v6i2.288
12
Guo, K., Tan, Z., Luo, E., & Zhou, X. (2021). Machine learning: The cyber-security, privacy,
and Public Safety Opportunities and challenges for emerging applications. Security and
Communication Networks, 2021, 1–2. https://doi.org/10.1155/2021/9870129
Jilcha Sileyew, K. (2020). Research design and methodology. Cyberspace.
https://doi.org/10.5772/intechopen.85731
Morfino, V., & Rampone, S. (2020). Towards near-real-time intrusion detection for IOT devices
using supervised learning and Apache Spark. Electronics, 9(3), 444.
https://doi.org/10.3390/electronics9030444
Patil, Varadarajan, V., Walimbe, D., Gulechha, S., Shenoy, S., Raina, A., & Kotecha, K. (2021).
Improving the Robustness of AI-Based Malware Detection Using Adversarial Machine
Learning. Algorithms, 14(10), 297–. https://doi.org/10.3390/a14100297
Rashid, M. M., Kamruzzaman, J., Hassan, M. M., Imam, T., & Gordon, S. (2020). Cyberattacks
Detection in IoT-Based Smart City Applications Using Machine Learning Techniques.
International Journal of Environmental Research and Public Health, 17(24), 9347–.
https://doi.org/10.3390/ijerph17249347
Shaukat, K., Luo, S., Varadharajan, V., Hameed, I. A., Chen, S., Liu, D., & Li, J. (2020).
Performance Comparison and Current Challenges of Using Machine Learning
Techniques in Cybersecurity. Energies (Basel), 13(10), 2509–.
https://doi.org/10.3390/en13102509
Sturm, Gerlacha, J., Pumplun, L., Mesbah, N., Peters, F., Tauchert, C., Nan, N., & Buxmann, P.
(2021). Coordinating human and machine learning for effective organization learning.
MIS Quarterly, 45(3), 1581–1602. https://doi.org/10.25300/MISQ/2021/16543
13
Vergne. (2020). Decentralized vs. Distributed Organization: Blockchain, Machine Learning and
the Future of the Digital Platform. Organization Theory, 1(4), 263178772097705–.
https://doi.org/10.1177/2631787720977052
Zhang, S., Xie, X., & Xu, Y. (2020). A Brute-Force Black-Box Method to Attack Machine
Learning-Based Systems in Cybersecurity. IEEE Access, 8, 128250–128263.
https://doi.org/10.1109/ACCESS.2020.3008433