Exploring the impact of AI-driven solutions on cybersecurity adoption in small and

medium enterprises
Geeta Sandeep Nadella *, Hari Gonaygunta, Deepak Kumar and Priyanka Pramod Pawar

Department of Information Technology, University of the Cumberlands, Williamsburg, Kentucky, USA.

World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

Publication history: Received on 07 March 2024; revised on 20 April 2024; accepted on 22 April 2024

Article DOI: https://doi.org/10.30574/wjarr.2024.22.1.1185

Abstract
This study delves into the intricate dynamics of cybersecurity adoption within small and medium enterprises (SMEs),
with a specific focus on the transformative potential of AI-driven solutions from a machine learning perspective.
Leveraging a rich and diverse dataset of real-world cybersecurity incidents tailored to the SME context, the research
meticulously investigates the efficacy of three prominent machine learning models: logistic regression, random forest,
and gradient boost classifiers. Through a comprehensive evaluation using a suite of performance metrics, including
accuracy, precision, recall, and F1 scores, the study scrutinizes the predictive capabilities of these models in forecasting
and categorizing cyber threats encountered by SMEs. The findings unveil nuanced insights into the multifaceted
challenges and opportunities inherent in the cybersecurity landscape of SMEs, shedding light on the complex interplay
between AI-driven solutions and evolving cyber threats. While the analysis reveals commendable performance across
the models, it also uncovers inherent limitations in accurately discerning and categorizing specific types of attacks.
These findings underscore the critical need for ongoing refinement and optimization of AI-driven cybersecurity
solutions, necessitating a continuous, iterative process informed by real-world data and adaptive learning mechanisms.
By harnessing the strengths of machine learning and embracing a proactive stance towards cybersecurity, SMEs can
fortify their defense mechanisms and safeguard their digital assets in the increasingly volatile and interconnected digital
ecosystem.

Keywords: Artificial Intelligence; Cyber Security; Machine Learning; Random Forest; Logistic Regression; Gradient
Boost; Small and Medium-Sized Enterprises (SMEs)

1. Introduction
The rapid proliferation of digital technologies and the ubiquity of internet-connected devices have ushered in an era of
unprecedented opportunities and vulnerabilities in the realm of cybersecurity [1]. Cybersecurity, a critical field of study,
safeguards organizations, interactions, systems, and information from a myriad of cyber-attacks [2]. The integration of
the Internet of Things (IoT) holds immense promise for innovative societal applications. However, it also heightens the
need for robust cybersecurity measures as financial institutions and other critical sectors face increasing cyber risks
[3]. In response to these evolving threats, researchers have proposed AI-driven cybersecurity approaches that leverage
advanced techniques, such as the K-NN algorithm combined with enhanced encryption and decryption procedures, to
mitigate network attacks that employ information poisoning and computational intelligence [4]. The incorporation of
artificial intelligence (AI) into cybersecurity solutions has emerged as a pivotal development, transforming the
landscape of cyber defense.

Historically, machine learning has been utilized in cybersecurity since the 1990s, but advancements in data and
computing have propelled AI to become a crucial component of modern cyberdefense strategies [5], [5]. Automated
cyber defenses can now mimic human intelligence and behavior, enabling faster detection of network vulnerabilities


Corresponding author: Dr. Geeta Sandeep Nadella
Copyright © 2024 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0.
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

compared to manual efforts [6]. The COVID-19 pandemic has further accelerated the adoption of technology, leading to
a rapid increase in cybercrime and putting businesses and individuals at risk [7]. Small and medium enterprises (SMEs)
are particularly vulnerable to the proliferation of cybercrime, as they often lack the resources and maturity to
implement robust cybersecurity measures [8]. According to recent estimates, the global cost of cybercrime is projected
to reach $10.5 trillion by 2025, highlighting the urgency for SMEs to address this growing threat [9]. Centralized
cybersecurity management and the availability of adequate IT resources have been shown to enhance an organization's
security posture, underscoring the need for SMEs to adopt a more strategic and proactive approach to cybersecurity
[10].

The evolving cybersecurity landscape and the increasing prominence of AI-driven solutions have prompted the research
community to investigate the factors that influence cybersecurity decision-making and adoption within SMEs. The
extended Technology-Organization-Environment (TOE) framework proposed by Wallace et al. [11] provides a
multidimensional perspective on the determinants of cybersecurity adoption, including new variables, practice
standards, cyber catalysts, and additional dimensions. Similarly, Hasan et al. [12] explored the relationship between
cybersecurity readiness characteristics and organizational performance using the TOE framework, highlighting the
importance of a comprehensive approach to cybersecurity management. This study aims to contribute to the growing
body of knowledge by exploring the impact of AI-driven solutions on cybersecurity adoption within small and medium
enterprises, leveraging a machine learning-based approach to uncover insights and inform future research and practical
applications.

2. Literature Review
Many cybercrimes, including hacking attacks and data breaches, have received extensive media coverage, elevating the
subject of cybersecurity to the forefront of public discourse. According to Zimmerman and Renaud's (2019)
investigation, rather than depending on technological means, they concentrated on improving aspects that lead to
resilient cybersecurity and positive outcomes [13]. According to Liu et al. (2020), a company's cybersecurity posture
improves when management is centralized [14]. Benz and Chatterjee (2020) noted that small and medium enterprises
(SMEs) are among the most fragile and immature business models, and the ever-increasing cyber hazards for SMEs are
a direct result of the proliferation of cybercrime, as elucidated [15], [16].

The availability and utilization of an organization's IT infrastructure are critical to the smooth running of its internal
business processes. Hasan et al. (2021) noted that when organizations assess the availability and utilization of IT, they
must consider the investment, capacity, and hardware components [17]. According to three studies, having enough IT
resources makes systems more secure and reduces the likelihood of security breaches and incidents [18]. Decisions
regarding the implementation of cybersecurity are significantly impacted by the revised and extended TOE framework
that was presented by Wallace et al. in 2020. This framework includes new variables, practice standards, cyber catalysts,
and extra dimensions under the classic organization, technology, and environment dimensions [19]. Hasan et al. (2021)
investigated the relationship between different cybersecurity readiness characteristics and organizational financial and
non-financial performance by using the TOE framework [17]. Brunner et al. (2021) assessed the state of risk
management techniques in connection with Information Security Management Systems (ISMS) for regions of Austria,
Germany, and Switzerland, examining how companies maintain data security by using standard operating procedures,
data-gathering methods, tool types, interpersonal arrangements, and documentation artifacts [20]. Cyber-physical
systems, also known as the CPS Environment, aim to create IoT infrastructure for educational research purposes in a
variety of Pes-related domains, with the primary objective of providing physical architecture and facilitating the
exploration of its possible practical uses by scholars and learners [21].

Studies have examined how investors perceive a company's likelihood of declaring bankruptcy in the future in relation
to eXtensible Business Reporting Language (XBRL). Standards for implementing XBRL reduce the expected risk of
crashes and companies with more volatile performance, opaque accounting, and untrustworthy projections from
analysts [22]. These problems likewise impact procedures used in Cyber Digit manufacturing (a DM), and creating a
meta-architecture might help resolve this problem [23]. A monetary risk related to cybersecurity is highlighted in a
hypothetical situation using the College of Industry and Computer Engineering's Explorers application, where the
evolutionary algorithm, or GA, is used to pick the components of the protective mechanism, and a fuzz assessment is
used to determine the degree of fitness [24].

1191
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

3. Methodology
The methodology employed in this research explores the impact of AI-driven solutions on cybersecurity adoption within
small and medium enterprises (SMEs) from a machine learning perspective. It begins with the collection of a dataset
sourced from cybersecurity attacks, tailored to SME contexts, followed by preprocessing steps to handle missing values
and ensure data quality. Descriptive statistics are computed to analyze numerical attributes, while exploratory data
analysis techniques uncover patterns and trends in the dataset, focusing on indicators of compromise and response
actions taken [25]. A temporal analysis is conducted to understand the dynamics of cyber threats over time. Predictive
models that leverage machine learning are developed to measure the efficacy of AI-driven solutions in mitigating cyber-
attacks.

In the machine learning modeling phase, three distinct algorithms - logistic regression, random forest, and gradient
boost - are employed to analyze the dataset and predict cyber threats within SMEs [26]. Logistic regression, a classic
statistical method, is utilized to model the probability of occurrence of a binary outcome, making it suitable for
classification tasks such as identifying malicious activities. Meanwhile, random forest, a powerful ensemble learning
technique, utilizes a multitude of decision trees to generate robust predictions by aggregating outputs of multiple
individual models [28]. Gradient boost, a cutting-edge boosting algorithm, revolutionizes cybersecurity modeling for
SMEs by iteratively refining predictive performance through sequential optimization [27].

Figure 1 Proposed Framework

The methodology employed in this research explores the impact of AI-driven solutions on cybers-security adoption
within small /medium enterprises (SMEs) from a machine learning perspective. It begins with the collection of a dataset
sourced from cybersecurity attacks, tailored to SME contexts, followed by preprocessing steps to handle missing values
and ensure data quality [28]. Descriptive statistics are computed to analyze numerical attributes, while exploratory data
analysis techniques uncover patterns and trends in the dataset, focusing on indicators of compromise and response
actions taken. A temporal analysis is conducted to understand the dynamics of cyber threats over time. Predictive
models that leverage machine learning are developed to measure the efficacy of AI-driven solutions in mitigating cyber-
attacks. Model performance is evaluated using appropriate metrics, and findings are interpreted to provide insights into
enhancing cybersecurity resilience in SME environments, accompanied by recommendations for future research and
practical implications [28].

3.1. Data Collection

In this research, data collection is paramount for exploring the impact of AI solutions on cybersecurity adoption in
enterprises of medium and small buildings. The dataset utilized is carefully curated from cybersecurity attacks,
explicitly focusing on incidents pertinent to SMEs. This dataset comprises various attributes crucial to understanding
cyber incidents, like the time stamp, protocol, packet length, packet type, source-destination IP addresses, source-port

1192
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

numbers for the destination, and Traffic Type, among others, all tailored to SME cybersecurity contexts. By
incorporating these attributes, the dataset enables a comprehensive analysis of cybersecurity threats faced by SMEs,
providing insights into the nature, frequency, and severity of cyberattacks targeting SE businesses. Through this data
collection process, the research lays the foundation for examining the efficacy of AI-driven solutions in bolstering
cybersecurity defenses within the SME sector, offering valuable insights to inform future cybersecurity strategies and
practices.

Figure 2 Cybersecurity Dataset

3.2. Data Preprocess

Information preprocessing is a vital step in preparing datasets for analysis, particularly in the context of exploring
impact-AI-driven-solutions on cyber-security adoption in small-medium enterprises from a machine learning
perspective [29]. Initially, the dataset undergoes preprocessing to address missing values and ensure completeness of
data. Columns such as Timestamp are converted to date time format to facilitate temporal analysis and enable
exploration of trends over time. Moreover, to convert categorical variables in a set of data into numbers that are
appropriate for algorithmic machine learning, methods of encoding, such as labeling and single-hot encoding, can be
applied. The algorithms for learning ML guarantee that it will work by allowing for the inclusion of categorical attributes
in analysis. Additionally, outlier detection techniques may be applied to identify and handle anomalies within data,
thereby enhancing data quality and robustness of subsequent analyses. Via accomplishment preprocessing stages, the
dataset is refined and optimized for more examination and modeling, facilitating a comprehensive investigation into AI-
driven solutions on cyber-security adoption in SMEs [30].

4. Data Analysis of Malicious Activity

In an analysis of malicious activity within small and medium enterprises (SMEs), the dataset is scrutinized to identify
patterns and characteristics indicative of cyber threats. This examination primarily focuses on various indicators of
compromise, such as IDS/IPS Alerts, Malware Indicators, and Alerts/Warnings. By delving into the indicators, the
research aims to understand the nature and severity of threats faced by SMEs in their cybersecurity landscape.
Additionally, attention is given to discerning top users and network segments associated with these malicious activities,
providing insights into potential vulnerabilities within SME networks. Through this analysis, a comprehensive
understanding of the threat landscape within SMEs is obtained, facilitating the development of effective cybersecurity
strategies and the adoption of AI-driven solutions to mitigate and counteract malicious activities.

1193
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

Figure 3 Attack Types Distribution

In Figure 3, the distribution of attack types within small and medium enterprises (SMEs) reveals intriguing insights into
prevailing cybersecurity threats. Among recorded attack types, disseminated denial attacks constitute a significant
portion, accounting for 12,500 instances. DDoS attacks, known for their disruptive nature, pose substantial risks to SMEs
by overwhelming their network infrastructure and rendering services inaccessible to legitimate users. Concurrently,
malware-based attacks totaling 12,300 instances represent another substantial threat vector. A wide range of harmful
software is used in virus assaults, with the goal of infiltrating networks, compromising the confidentiality of
information, and interfering with corporate processes. Intrusion attempts, comprising 13,000 instances, constitute a
notable portion of the attack landscape. Intrusions involve unauthorized access attempts into network systems,
potentially leading to data breaches, system compromise, and unauthorized data access.

Figure 4 All network segments in SMEs

In Figure 4, the proportion of severity levels across all network segments showcases a balanced distribution, with each
severity level constituting approximately one-third of the total distribution. Specifically, analysis reveals that 33% of
incidents are categorized as low severity, 34% as high severity, and another 33% as medium severity. This equilibrium
suggests a diverse range of cybersecurity incidents occurring across SME network segments, with varying degrees of
impact and criticality. The prevalence of high-severity incidents highlights the potential for significant disruptions and
damage to SME operations and assets.

4.1. Machine Learning Modeling

In the machine learning modeling phase, three distinct algorithms—logistic regression, random forest, and gradient
boost—are employed to analyze datasets and predict cyber threats within small and medium enterprises (SMEs).
Logistic regressions, a classic arithmetical method, is utilized to model the phase likelihood of occurrence of a binary
outcome, making it suitable for classification tasks such as identifying malicious activities. Meanwhile, random forest, a
powerful collaborative knowledge technique, utilizes a multitude of decision trees to generate robust predictions by
aggregating outputs of multiple individual models.

1194
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

Logistic Regression: In the background of cybersecurity analysis for small and medium enterprises, logistic regression
serves as a foundational model for predicting and classifying cyber threats. Because of its simplicity and interpretability,
logistic regression is used to estimate the probability of a binary outcome, making it suitable for identifying malicious
activities within SME networks. By analyzing various features and attributes from the dataset, logistic regression
computes the likelihood of specific cyber events occurring, providing valuable insights into potential vulnerabilities and
threat vectors. Despite its linear nature, logistic regression remains a valuable tool in the cybersecurity arsenal,
particularly for scenarios where transparency and explainability are paramount.

Random Forest: Random forest, a versatile ensemble learning technique, offers SMEs a robust approach to cybersecurity
modeling by harnessing the power of multiple decision trees. Individually, the D-tree within the random-forests
ensemble autonomously learns from different subsets of the dataset, capturing unique patterns and relationships
between features. Through aggregation of predictions from individual trees, random forest generates highly accurate
and reliable forecasts while mitigating the risk of overfitting. Logistic regression is particularly effective in handling the
complex and heterogeneous data characteristic of cybersecurity incidents in SMEs. Leveraging the shared
understanding of manifold decision trees with random forests, SMEs can discern subtle nuances in cyber threat
landscapes and proactively fortify their defense mechanisms.

Gradient Boost: Gradient boost, a cutting-edge boosting algorithm, revolutionizes cybersecurity modeling for SMEs by
iteratively refining predictive performance through sequential optimization. By constructing a sequence of weak
learners—typically decision trees—gradient boost continuously focuses on minimizing prediction errors, gradually
improving the model correctness with apiece repetition. This iterative learning procedure enables gradient boost to
imprisonment multifaceted connections plus non-linear relationships within the dataset, leading to enhanced predictive
capabilities. Additionally, gradient boost adapts its focus to prioritize misclassified instances, allowing SMEs to
effectively allocate resources towards addressing the most critical cyber threats. Through its adaptive and data-driven
approach, gradient boost empowers SMEs to stay ahead of evolving cyber threats and bolster their cybersecurity
defenses with precision and efficacy.

4.2. Results Evaluation

In the results evaluation phase, the performance of logistic regression, random forest, and gradient boost models is
meticulously assessed to ascertain their effectiveness in predicting and mitigating cyber threats within small and
medium enterprises (SMEs). The predictive capability and generalization proficiency of these algorithms are
quantitatively evaluated using a comprehensive suite of metrics, including precision, recall, recall accuracy, F1 scores,
and the area under the receiver operating characteristic curve (AUC-ROC). By juxtaposing these metrics across different
models, businesses can derive nuanced insights into the relative advantages and disadvantages, as well as the potential
trade-offs and challenges, associated with each approach. This analytical framework facilitates informed decision-
making by highlighting the most productive strategies for enhancing cybersecurity postures in the context of SMEs.

Table 1 Models Comparison results

Models-names Attack-Type Accuracy Precisions Recall F1-S

Logistic Regression DDoS 0.34 0.34 0.48 0.39
Intrusion 0.33 0.19 0.24
Malware 0.34 0.34 0.34
Random Forest Classifier DDoS 0.33 0.32 0.36 0.34
Intrusion 0.33 0.31 0.32
Malware 0.32 0.31 0.32
Gradient Boosting Classifier DDoS 0.33 0.33 0.41 0.37
Intrusion 0.34 0.26 0.29
Malware 0.33 0.33 0.33

The main results of classification models (Logistic Reg, R-Forest, and Gradient-Boost) reveal similar overall
performance, with slight variations in precision, recall, and F1-score across different classes. Logistics regression attains
an accuracy of 34%, while the Random Forest classifier and gradient-boosting classifier exhibit slightly lower accuracies

1195
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

of 33.5% and 33%, respectively. Upon closer examination, it becomes apparent that the models struggle to classify
instances accurately across all classes, as evidenced by the relatively low precision, recall, and F1 scores for each class.
Specifically, the models demonstrate challenges in distinguishing between dissimilar kinds of fake dangers, such as
DDoS attacks, Intrusions, and Malware infections. Despite these limitations, the models achieve a balanced performance
across the classes, as indicated by the similar macro and weighted average metrics.

5. Conclusion
In conclusion, the evaluation of logistic regression, random forest classifiers, and gradient boosting are these models
implemented for predicting cyber threats within small and medium enterprises (SMEs), which underscores both the
potential and challenges of utilizing machine learning in cybersecurity. While all three models exhibit comparable
overall accuracies, precision, recall, and F1-scores, they demonstrate limitations in accurately classifying specific types
of attacks such as DDoS, Intrusions, and Malware. Despite these challenges, the models provide valuable insights into
the cybersecurity landscape of SMEs and serve as foundational tools for threat detection and mitigation. Moving
forward, further optimization and refinement of these models, coupled with ongoing checking and variation near
developing dangers, are indispensable to ensure the efficiency of AI-driven cybersecurity solutions in safeguarding SME
networks. Additionally, the findings emphasize the importance of a holistic approach to cybersecurity, integrating
advanced machine learning techniques with robust security protocols and practices tailored to the unique needs and
constraints of SME environments. By leveraging the strengths of machine learning and continuously improving
cybersecurity strategies, SMEs can better safeguard their electronic resources and reduce the dangers associated with
cyberattacks in the networked world of tomorrow.

Compliance with ethical standards

Disclosure of conflict of interest

No conflict of interest to be disclosed.

References
[1] National Academies of Sciences Engineering, and Medicine, Intelligence Community Studies Board, Computer
Science and Telecommunications Board, & division on Engineering and Physical Sciences A. Johnson, E.
Grumbling (Eds.), Implications of artificial intelligence for cybersecurity: Proceedings of a workshop, National
Academies Press (2019), 10.17226/25488
[2] Wiafe, F.N. Koranteng, E.N. Obeng, N. Assyne, A. Wiafe, S.R. Gulliver Artificial intelligence for cybersecurity: A
systematic mapping of literature IEEE Access, 8 (2020), pp. 146598-146612
[3] Y. Li, Q. Liu A comprehensive review study of cyber-attacks and cyber security; Emerging trends and recent
developments Energy Reports, 7 (2021), pp. 8176-8186
[4] X. Yu, H. Guo A survey on IIoT security 2019 IEEE VTS Asia Pacific wireless communications symposium (APWCS)
(2019), pp. 1-5
[5] Kaplan, M. Haenlein Siri, Siri, in my hand: Who's the fairest in the land? On the interpretations, illustrations, and
implications of artificial intelligence Business Horizons, 62 (1) (2019), pp. 15-25
[6] Z. Zhang, H. Ning, F. Shi, F. Farha, Y. Xu, J. Xu, …, K.K.R. Choo Artificial intelligence in cybersecurity: Research
advances, challenges, and opportunities Artificial Intelligence Review (2021), pp. 1-25
[7] I.C. Eian, L.K. Yong, M.Y.X. Li, YH Qi, Z. Fatima Cyber-attacks in the era of COVID-19 and possible solution domains
(2020)
[8] J. Scott, M. Kyobe Trends in cybersecurity management issues related to human behaviour and machine learning
2021 International Conference on Electrical, computer and energy technologies (ICECET), IEEE (2021), pp. 1-8
[9] M. Wazid, A.K. Das, V. Chamola, Y. Park Uniting cyber security and machine learning: Advantages, challenges and
future research ICT express, Korean Institute of Communication Sciences (2022), pp. 313-321,
10.1016/j.icte.2022.04.007 Vol. 8, Issue 3
[10] Zimmermann V, Renaud K (2019) Moving from a 'human-as-problem" to a 'human-as-solution" cybersecurity
mindset. International Journal of Human-Computer Studies 131: 169–187

1196
 World Journal of Advanced Research and Reviews, 2024, 22(01), 1190–1197

[11] Li Y, Liu Q (2021) A comprehensive review study of cyber-attacks and cyber security; Emerging trends and recent
developments. Energy Rep 7:8176–8186
[12] Benz M, Chatterjee D (2020) Calculated risk? A cybersecurity evaluation tool for SMEs. Bus Horiz 63(4):531–540
[13] Simola J (2019) Comparative research of cybersecurity information sharing models. Inf Secur: Int J 43(2):175–
195.
[14] Hasan S, Ali M, KurniaThurasamy SR (2021) Evaluating the cyber security readiness oforganizations and its
influence on performance. J Inf Secur Appl 58:102726
[15] Wallace S, Green KY, Johnson C, Cooper J, Gilstrap C (2020) An extended TOE framework forcybersecurity-
adoption decisions. Commun Assoc Inf Syst 47:338363.
[16] LAZIĆ, L. Benefit from AI in cybersecurity. In Proceedings of the 11th International Conference on Business
Information Security (BISEC 2019), Belgrade, Serbia, 18 October 2019. [Google Scholar]
[17] How Credit Card Companies are Fighting Cyber Frauds. Available
online: https://cio.economictimes.indiatimes.com/news/digital-security/heres-how-visa-
mastercard-and-paypal-are-fighting-cyber-frauds-with-ai/79381050 (accessed on 3 March 2021).
[18] Vähäkainu, P.; Lehto, M. Artificial intelligence in the cyber security environment. In Proceedings of the ICCWS
2019 14th International Conference on Cyber WarfarSe and Security: ICCWS, Stellenbosch, South Africa, 28
February–1 March 2019; p. 431.
[19] Amazon Web Services, Inc. Amazon Macie F.A.Q. Amazon. 2018. Available
online: https://aws.amazon.com/macie/faq (accessed on 3 March 2021).
[20] Proko, E.; Hyso, A.; Gjylapi, D. Machine Learning Algorithms in Cyber Security. In RTA-CSIT; 2018; pp. 203–207.
Available online: https://www.semanticscholar.org/paper/Machine-Learning-Algorithms-in-Cyber-
Security-Proko-Hyso/67525df429c50af9ae5fe10949cd7d279ee1184f (accessed on 27 October 2021).
[21] Orche, A.E.; Bahaj, M. Approach to Combine an Ontology-Based on Payment System with Neural Network for
Transaction Fraud Detection. Available online: https://astesj.com/v05/i02/p69/ (accessed on 27 October
2021).
[22] Tech Giants Using AI against Hackers. Available online: https://analyticsindiamag.com/how-tech-
giants-like-amazon-microsoft-google-are-using-ai-against-hackers/ (accessed on 3 March 2021).
[23] Hackers Trick Tesla. Available
online: https://www.technologyreview.com/2019/04/01/65915/hackers-trick-teslas-autopilot-
into-veering-towards-oncoming-traffic/ (accessed on 3 March 2021).
[24] Ford, V.; Siraj, A. Applications of machine learning in cyber security. In Proceedings of the 27th International
Conference on Computer Applications in Industry and Engineering, New Orleans, LA, USA, 13–15 October 2014;
IEEE Xplore: Kota Kinabalu, Malaysia, 2014; Volume 118. [Google Scholar]
[25] Jayal, A.; McRobert, A.; Oatley, G.; O'Donoghue, P. Sports Analytics: Analysis, Visualisation and Decision
Making in Sports Performance; Routledge: Abingdon, UK, 2018. [Google Scholar]
[26] UK Small Business Statistics, FSB. The Federation of Small Businesses. Available
online: https://www.fsb.org.uk/uk-small-business-statistics.html (accessed on 1 September 2021).
[27] SME Action Plan. Available
online: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/961722/SME-Action-Plan.pdf (accessed on 1 September 2021).
[28] The Impact of the Coronavirus so Far: The Industries that Struggled or Recovered—Office for National Statistics
(Ons.Gov.UK). Available
online: https://www.ons.gov.uk/economy/economicoutputandproductivity/output/articles/theimpa
ctofthecoronavirussofartheindustriesthatstruggledorrecovered/2020-12-09 (accessed on 1 September
2021).
[29] O'Leary, D.E.' Big Data', the 'Internet of Things', and the 'Internet of Signs.' Intell. Syst. Account. Financ.
Manag. 2013, 20, 53–65.
[30] Kirby, M.; Konbel, F.; Barter, J.; Hope, T.; Kirton, D.; Madry, N.; Manning, P.; Trigges, K. Sociology in Perspective;
Heinnemann: Oxford, UK, 2000.

1197