Computers and Electrical Engineering 110 (2023) 108813

Contents lists available at ScienceDirect

Computers and Electrical Engineering

journal homepage: www.elsevier.com/locate/compeleceng

Deep learning technology of computer network security detection

based on artificial intelligence
Qinghui Liu a, Tianping Zhang b, *
a
Network Center, Zaozhuang University, Zaozhuang 277160, Shandong, China
b
School of Communication and Information Engineering, Shanghai Technical Institute of Electronics & Information, Shanghai 201411, China

A R T I C L E I N F O A B S T R A C T

This paper is for CAEE special section VSI: With the continuous progress of the information age and science and technology, how to combine
aismg. Reviews processed and recommended complex and changeable data with self-protection ability has become an urgent problem that
for publication to the Editor-in-Chief by Guest needs to be solved. In the computer system, the traditional technology for protection is somewhat
Editor Dr J Dinesh Peter
passive. To this end, this paper analyzes computer network security monitoring under artificial
intelligence (AI) deep learning technology and discusses its development trend, which monitors
Keywords:
network security to find and address problems in time. The results show that the deep learning
Network security monitoring
Deep learning technology of computer network security detection can improve its security performance with
Artificial intelligence high safety performance, high detection rate, and low false alarm rate. It can monitor the vul­
Information science nerabilities in the computer network in time and realize effective security attack detection of the
Computer network computer network.

1. Introduction

Along with the advancement of computer network technology, there is a growing awareness of AI and computer security. Computer
network security refers to protecting the security of human-centered information systems and ensuring that data and information are
not illegally obtained and manipulated. Therefore, people have carried out in-depth investigation and exploration and have become
the current research focus.
Due to the rising awareness of security and privacy, many scholars have conducted research on them. Sulicdio Julias believed that
computer security was an essential factor in the information technology world. Negatively, there were more types of intrusion or attack
in the network. Therefore, it is necessary to determine the interference or threat in the network to reduce intrusion [1]. Tang Ying
analyzed a simulation model for computer network security evaluation based on neural networks. First, based on the system security
experimental platform, a system security detection algorithm was built, and the strategy for selecting the computational scale was
analyzed [2]. Kebande Victor R investigated the pertinence of real-time monitoring (RTM) as a vigilant and complementary security
component in modern network environments, in particular for appropriate planning, preparation and alleviation in the event of a
cyber security incident [3]. Alabady Salah A believed that with the increasing number of connected devices in the Internet of Things,
the threats to confidentiality and security were also growing [4]. Choi Seul-Ki’s research found that computer devices lacking security
features were compromised by assailants, leading to many security breaches. In particular, an assailant could take control of a

This paper was recommended for publication by Associate Editor: Dr J Dinesh Peter
* Corresponding author.
E-mail addresses: [email protected] (Q. Liu), [email protected] (T. Zhang).

https://doi.org/10.1016/j.compeleceng.2023.108813
Received 22 February 2023; Received in revised form 5 June 2023; Accepted 7 June 2023
Available online 14 June 2023
0045-7906/© 2023 Elsevier Ltd. All rights reserved.
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

computer device with insufficient security features [5]. Shone Nathan found that the main task of the network in the future was to
establish an intelligent network architecture as far as possible to achieve intelligence. Software-defined network (SDN) technology has
broken the tight coupling between the control plane and data plane in traditional network architecture, which has made the
controllability, security and economy of network resources a reality [6]. The above are the relevant research on computer network
security monitoring. Times have moved on, and new technologies have emerged to be used in computer network security monitoring.
An increasing number of artificial intelligence techniques are being applied to computer network security detection, so many
academics have researched and analyzed this. Sarker Iqbal H found that DL derived from artificial neural networks was one of the main
techniques for intelligent network security systems or policies to operate in an intelligent manner today [7]. Zhang Jun’s research
found that with the vigorous development of network attacks and cyber crimes against network physical systems, the detection of these
attacks was still challenging. DL provided better performance than machine learning because of its hierarchical settings and effective
algorithms for extracting useful information from training data. The deep learning model was rapidly used for network attacks against
physical systems [8]. Li Jian-hua believed that there was a broad interdisciplinary intersection between network security and AI.
Therefore, the AI model needs specific network security defense and protection technologies to combat adversarial machine learning
and maintain privacy in machine learning to ensure the security of joint learning [9]. Ahsan Mostofa found that machine learning was
of rising importance in cybersecurity. The main aim of machine learning with the utilization of artificial intelligence in cybersecurity
was to make the malware detection process more actionable, extensible and valid than conventional methods because traditional
methods required human intervention, and the network security field involved challenges of AI, which required efficient methods and
theoretical processing [10]. Zhao Shanshan found that technology based on computational intelligence played a key role in addressing
the network security challenges of complex systems such as the Internet of Things and Cyber Physical System. At present, the Internet
of Things is facing increasingly more security problems [11]. Zhang Zhimin examined the use of AI techniques in a broad range of
cybersecurity applications and identified some limitations and challenges. Based on his findings, he proposed a conceptual
human-in-the-loop intelligent cybersecurity model [12]. The above are about the relevant research of AI in network security
monitoring.
With the rapid advancement of science and technology, the issue of cybersecurity has become increasingly important. The com­
puter network system has been attacked by various viruses or hackers during its operation. Therefore, it is of great significance to
effectively prevent these harmful events, such as viruses and hacker attacks, from threatening computers and protecting users’ in­
formation property from being infringed upon. This study analyzes the method of computer network security monitoring, analyzes the
method through experimental analysis and draws conclusions.
This paper mainly introduces the research status and existing problems of artificial intelligence technology in computer network
security detection. From the perspective of artificial intelligence learning and research, it proposes a computer security detection
method based on a deep neural network and applies it to the field of computer network security.
The first part first studies the research background and references of computer network security detection. The second part in­
troduces the relevant theories. This paper introduces computer network security and artificial intelligence. The third part has the
computer network security monitoring and management of artificial intelligence, which mainly includes the introduction of the al­
gorithm. The fourth part presents the experimental findings. We set the routine computer network security monitoring record as a
control group, set the computer network security monitoring record under AI deep learning technology as the experimental group and
let them perform the control experiments. It was found that the false alarm rate of computer network security under AI deep learning
technology, had better performance, and a higher safety profile.

2. Computer network security detection of DL algorithm

2.1. Computer network security

2.1.1. Computer virus

Computer viruses are destructive data and functions introduced into computer programs that affect the normal operation of
computer networks. Computer viruses are very destructive and spread very quickly, which is one of the main reasons for network
security problems. Computer viruses are highly infectious and can replicate themselves in the process of transmission, leading to the
collapse of the network system and even causing serious economic losses [13]. At the same time, computer viruses are hidden, latent
and difficult to eradicate, which poses a serious threat to the security of computer networks.

2.1.2. Physical security

Physical security issues refer to natural disasters, such as earthquakes, floods, fires, etc. These disasters make computer networks
unable to operate normally or even destroyed [14]. The network may fail or even cause damage to computer equipment. Electro­
magnetic interference and lightning can also damage computer networks. Therefore, in the process of using the computer network,
attention should be given to the physical security of the computer, and active measures should be taken to avoid electromagnetic
interference, lightning, fire and floods.

2.1.3. System vulnerabilities

System vulnerabilities are security risks of computer hardware, software and protocols. Users can gain unauthorised access to
computer systems. Although computer systems are constantly updated with the development of science and technology, system
vulnerabilities are inevitable. Many computer programs and systems are designed to be vulnerable. These vulnerabilities are the

2
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

Fig. 1. Framework diagram for computer network security issues.

targets of hackers and computer viruses. To ensure the security of computer networks, system vulnerabilities must be eliminated in
time.

2.1.4. Man-made attack

Hacker attack refers to the behavior in which hackers take advantage of vulnerabilities in computer systems to undermine the
security of computer networks to obtain network data [15,16]. Hacker attacks can be extremely destructive because many hackers
steal personal data, user information, trade secrets, etc., for economic benefits. People cause huge economic losses and even endanger
social security. Second, there are malicious and man-made attacks on the internet, which use computer vulnerabilities to destroy the
integrity of network information and even access sensitive information without affecting the normal operation of the network. For a
more detailed presentation, this is illustrated in Fig. 1.

2.2. Overview of AI

The discussion of the concepts and current status of AI development helps technicians gain theoretical insight into the dimensions
and provides a solid basis for the further development and application of the relevant technologies. Artificial intelligence, based on
computer technology, can quickly, efficiently and accurately process information and data by simulating human thinking. With the
continuous growth of AI technology, it has been widely used in various fields of production and life, making contributions to economic
development and social life and responding to the needs of actual use. AI technology is based on a computer framework, with software
algorithms and hardware support to simulate human thinking [17,18]. This is a feature of AI technology, which enables it to solve
complex problems. The application framework of AI technology is shown in Fig. 2.
The advantages of AI include the following:

2.2.1. Strong learning ability

AI can learn like human beings and has a unique learning mechanism. Different from the human learning mode, AI has the ability to
learn independently without human intervention and can independently extract valuable information from a huge database.

2.2.2. Strong computing ability

Computers are now tens of millions of times faster than humans. With the powerful computing power of computers, AI can process
more fuzzy information and obtain relatively complete information to complete tasks more effectively, thus significantly improving
work efficiency [19].

2.2.3. Reduction of resource consumption

With the help of AI, the intelligent control of resource consumption can be realized, resulting in huge resource savings, significant
cost reduction, significant energy consumption reduction, and significant improvement of work efficiency.

2.3. Application of AI technology in network security

AI has a wider application direction in network security management [20,21]. It can effectively improve the firewall capability of
the computer itself to improve the computer defense capability, such as computer network viruses. AI technology can be used to

3
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

Fig. 2. Framework diagram for the application of AI technologies.

Fig. 3. AI technology interception schematic.

intelligently prevent spam and monitor and intercept intrusion. Once statistics and memory are used for defense, the characteristics of
attackers and spammers can be transferred to their own databases and shared, thus establishing global protection against viruses.
When people browse hardware and download software from the internet, these software and files may contain some viruses and
advertisements. At this time, AI technology can actively monitor and scan software packages or files to ensure their safety. If a

4
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

problematic installation program or software package is detected, the computer uses administrator rights to block and isolate it and
then remind the user to minimize the threat of downloading files or software. The diagram of risk factor interception under AI is shown
in Fig. 3.
The application of AI in computer network security management mainly includes data mining technology, memory and learning.
The characteristics of computer and network connections are mainly obtained by processes such as review and aggregation. However,
AI technology does not interfere with the normal network activities of computers. Once an exception is detected in a computer network
activity, it reports the exception to the user. Once it exceeds its memory, it starts the computer firewall to forcibly block the computer
to forcibly intercept, diagnose and repair abnormal data. To better address the new situation in the process of computer network
management, in the AI control link, the monitoring and alarm module should be optimized. Each operation link of the computer
network is monitored in real time to ensure that once problems occur in the operation process, AI control technology can be found early
and quickly. AI control technology can also be configured with the use authority. Different administrators have a certain degree of
difference in their understanding of management level and job responsibilities. Therefore, the use of management authority can avoid
the risk caused by management errors and facilitate the management of computer network managers to realize the implementation of
management responsibilities. Artificial intelligence and computer network technology are complementary and interdependent, and
the security of computer networks can be guaranteed through the rational use of artificial intelligence technology. Therefore, it is
necessary to actively promote the function of AI in the application of computer network technology and promote its best use.

2.4. Application of DL technology in network security

Cyber security is a nonexistent psychological game between the protector and the attacker [22,23]. In this kind of psychological
game, the attackers are motivated by various technologies and programs, while the defenders of the game are usually relatively
passive. One way to change the rules of the game is to use the combination of human and machine intelligence, rather than hard coded
control mechanisms, to defend against various attacks, especially complex attacks. An essential representation of this intelligence is
machine learning. Traditional machine learning includes finding a feature vector (that is, a group of features) based on the problem to
be solved, extracting training data, and building a training model based on the feature vector. In many cases, the quality of features not
only affects the quality of training but also affects the efficiency of training. Usually, a set of feature vectors with semantic overlap are
selected. These eigenvectors have a high computational dimension and do not increase the actual performance. One approach to
solving this issue is to reduce the dimensionality using methods such as principal component analysis. The other is to use DL, which has
become very common in recent years.
The DL method helps to analyze data representations at multiple levels. DL methods focus on problems to identify feature vectors
and process data by learning and extraction. For example, in a botnet, a model using its relevant features is needed to identify
completely and effectively. In this method, the quality of the feature must first be determined. The quality of features directly de­
termines the effectiveness of the final learning results, which usually requires the use of cross-validation vectors presented in se­
mantics. Some focus on system discovery features, and some focus on establishing nonlinear models, so there are great differences in
different fields. Because of its use in various fields, it has different study priorities. In terms of computer network security, the most
critical issue is security requirements, which must be considered according to technical development and economic exchange. If the
actual computer network does not meet the corresponding security requirements, it cannot provide users with corresponding services
in formal operation, which reduces the operating efficiency of the entire computer system. Computer network security has strong
integrity. Therefore, the analysis must be carried out from a number of viewpoints to ensure the validity of the final outcome of the
analysis. The above analysis shows that the security of computer networks can be improved based on the DL algorithm. Moreover,
computer network security can be analyzed from all aspects to improve the security of computer network operation.

3. Computer network security monitoring management of AI

AI is a subdivision of computer science that is used to study machine intelligence and develop intelligent machines or intelligent
systems. Artificial methods and technologies are used to imitate, expand and enhance human intelligence. The combination of AI
technology and computer network security monitoring has brought many scientific research achievements. Intrusion detection is a
kind of active defense that collects information from key computer networks and host locations and is analyzed to find signs of security
policy violations or intrusions. Appropriate actions are taken to take the initiative. Therefore, this article chooses a fuzzy clustering
algorithm based on AI to monitor computer network security to increase the detection rate and decrease the false alarm rate.

3.1. Factors affecting computer security

To address issues in the area of computer cybersecurity, there is a need to learn DL methods based on image data and to redesign
discrete data processing. For statistical measurement and test calculation, the statistical data used in the hypothesis test of the pro­
portion of two independent samples are:
∑m
oj = − k i=1 fij lnfij (1)

The following can be obtained:

5
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

/ ( )
∑m ∑n ∑m
Vj = 1 + k i=1
fij lnfij j=1
1+k i=1
f lnfij (2)

Data are replaced for calculation:

/
∑n
Vj = bj j=1
bj (3)

3.2. Clustering analysis algorithm model

Cluster analysis, one of the approaches to multivariate statistical analysis, is an important area of unsupervised pattern classifi­
cation in statistical pattern recognition and a process of distinguishing and classifying things according to certain requirements and
rules [24,25]. In this process, an absence of prior knowledge of classification and no guidance is given. The similarity of things is
completely relied on as the classification standard, so it is an unsupervised classification.
Let A = {a1a2,..., an} be the sample set waiting for clustering.ai = (ai1,ai2,..., aim) represents the m characteristic values of the ith
sample. Therefore, the following conditions are needed:
A1 ∪ A2 ∪ ... ∪ Ak = A (4)

Ai ∩ Aj = A⊗, 1 ≤ i ∕
=j≤k (5)

The affiliation of sample ai(1 ≤ l ≤ m) to a subsection can be represented by the function:

{
1, al ∈ Ai
μAi (al ) = μil = (6)
0, al ∕
∈ Ai
{ }
j
∑
Ng = μil |μil ∈ {0, 1}; μil = 1 (7)
j=1

Among them, the affiliation function must satisfy condition μil ∈ Ng. This means that each sample must be of a single class and that
each subsection cannot be empty, which is often referred to as hard partitioning.
The hard division can only take values on the k unit vectors of the unit hypercube, namely:
{ ⃒ }
Ng = μj ∈ Ng ⃒μji ∈ {0, 1}, ∀i (8)

In fuzzy clustering, the membership function μil of the sample is expanded from binary value {0, 1} to interval [0, 1], and the
condition μil ∈ Nf is met. The following results are obtained:
∑
k
Nf = {μil |μil ∈ [0, 1]; μil = 1, ∀l} (9)
j=1

The fuzzy division occurs on the hyperplane of the valence unit basis vector of the value range, namely:
⃒
{ ⃒∑ k
⃒
Nf = μj ∈ Nf ⃒ μ = 1, ∀i} (10)
⃒ j=1 ji

For a given dataset, fuzzy clustering analysis can be used to easily obtain the fuzzy K partition. For a dataset with real charac­
teristics, namely, A⊂Tn, there are:

B2 (al , qi ) = (al − qi )R (al − qi ) (11)

To ensure that the clustering results can make “birds of a feather flock together”, the distortion degree of each class is used to
construct the fuzzy clustering objective function. Then, the objective function is optimized to obtain the optimal fuzzy set of the largest
sample set. Common forms of fuzzy objective functions are as follows:
∑
k ∑
m
S(U, Q) = μnil [B(al , qi )]2 , μil ∈ [0, 1] (12)
i=1 l=1

Among them, parameter n is the fuzzy degree control factor.

4. Computer network security experiment under AI DL technology

4.1. Detection rate

AI deep learning technology can detect vulnerabilities in computer networks in time and realize effective security attack detection

6
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

Fig. 4. Computer network security detection rate analysis.

Fig. 5. Computer network security false alarm rate analysis.

on computer networks. Therefore, conventional computer network security monitoring is recorded as the control group, and computer
network security monitoring under AI depth learning technology is recorded as the experimental group. Their detection rates in seven
computer test sets are analyzed, as shown in Fig. 4.
As shown in Fig. 4, the detection rate of computer network security detection based on AI is higher than that of traditional computer
network security detection. The detection rate of computer network security based on AI is more than 80%. However, the conventional
computer security detection rate fluctuates greatly and is not stable, and the detection rate is low. Therefore, the vulnerabilities in the

7
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

Fig. 6. Analysis of the security performance of computer networks.

computer network cannot be found in time. The average detection rate of computer cybersecurity based on AI in the seven test sets is
approximately 87.9% and that of conventional computer security in the seven test sets is approximately 62.7%. There is a certain gap
between the two detection rates.

4.2. False alarm rate

False alarms may occur in computer network security detection, which affects the work efficiency and delays the computer work
process. Therefore, this article analyzes the false alarm rate of the computer test set for the control group and the experimental group,
as shown in Fig. 5.
Fig. 5 shows that the false alarm rate of computer network security detection based on AI is lower than that of traditional computer
network security detection. The lowest false alarm rate of computer cybersecurity detection based on AI is only 5%, but the highest
false alarm rate of traditional computer security detection is 31%. The difference between the two is relatively large. The average false
alarm rate of computer cybersecurity detection based on AI in seven test sets is approximately 12% and that of traditional computer
network security detection in the seven test sets is approximately 23%. In conclusion, computer network security detection based on AI
is feasible.

4.3. Safety performance

The ultimate goal of research on the DL technology of computer cybersecurity detection based on AI is to protect the security
performance of computer networks. Therefore, this article analyzes the security performance of the computer test set for the control
group and the experimental group, as shown in Fig. 6.
Fig. 6 shows that the computer cybersecurity performance index based on AI is higher than the conventional computer cyberse­
curity performance index. The computer cybersecurity performance index based on AI is more than 80, but the traditional computer
security performance index is only approximately 60. The difference between the two is relatively large. The average security per­
formance index of computer cybersecurity detection based on AI in the seven test sets is approximately 86.14. The average security
performance index of traditional computer network security detection in the seven test sets is approximately 66.14, which is
approximately 20 higher. In conclusion, computer network security detection based on AI is feasible.

4.4. Effectiveness

In this paper, the control group and the experimental group are analyzed for the effectiveness of the computer test set, as shown in
Fig. 7.
Fig. 7 shows that the effectiveness of computer network security detection based on AI is higher than that of traditional computer
cybersecurity performance detection. The effectiveness of computer cybersecurity detection based on AI is between 75 and 95, and the
difference between them is relatively large. The effectiveness index of computer cybersecurity detection based on AI in seven test sets is

8
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

Fig. 7. Analysis of the effectiveness of computer network security.

Fig. 8. Comprehensive evaluation analysis of computer network security testing.

approximately 84.57 on average. The effectiveness index of conventional computer network security detection in seven test sets is
approximately 62.43 on average. Therefore, computer network security detection based on AI is feasible.

4.5. Comprehensive evaluation

This paper conducts a comprehensive evaluation on this through weighted evaluation, as shown in Fig. 8. From Fig. 8, the
comprehensive evaluation of the experimental group is higher and more feasible.

5. Conclusion

The development and application of artificial intelligence in computer network technology has greatly improved the security of
computer networks and effectively prevented and solved various risks. At the same time, artificial intelligence technology promotes
the intelligence of computer networks and improves the practicability of artificial intelligence technology. Artificial intelligence is a
complex and multidisciplinary technology with a variety of intelligent theories and methods. Therefore, researchers need to have some
professional basic knowledge of computer network security detection. Using the deep learning technology of artificial intelligence,

9
Q. Liu and T. Zhang Computers and Electrical Engineering 110 (2023) 108813

computer network security monitoring can be found. Compared with conventional computer network security monitoring, for
different computer monitoring methods, the false alarm rate decreases, improving the security of the network. At the same time, the
effectiveness of technology monitoring is greatly improved because the comprehensive evaluation analysis of the security test is also
higher. Therefore, the overall performance of computer network security monitoring using AI deep learning technology is better.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Data availability

No data was used for the research described in the article.

References

[1] Sulicdio Julias, Umi Kalsum Toibah, Arliando Yode. Comparative Analysis of Wireshark and Windump Software in Network Security Monitoring. J Media Comp
Sci 2022;1.1:1–6.
[2] Tang Ying, Elhoseny Mohamed. Computer network security evaluation simulation model based on neural network. J Intellig Fuzzy Syst 2019;37.3:3197–204.
[3] Kebande Victor R, Karie Nickson M, Ikuesan Richard A. Real-time monitoring as a supplementary security component of vigilantism in modern network
environments. Int J Inform Tech 2021;13.1:5–17.
[4] Alabady Salah A, Al-Turjman Fadi, Din Sadia. A novel security model for cooperative virtual networks in the IoT era. Int J Parallel Program 2020;48.2:280–95.
[5] Choi Seul-Ki, Yang Chung-Huang, Kwak Jin. System hardening and security monitoring for IoT devices to mitigate IoT security vulnerabilities and threats. KSII
Trans Internet Inform Syst (TIIS) 2018;12.2:906–18.
[6] Shone Nathan. A deep learning approach to network intrusion detection. IEEE Trans Emerg Topics Comput Intellig 2018;2.1:41–50.
[7] Sarker Iqbal H. Deep cybersecurity: a comprehensive overview from neural network and deep learning perspective. SN Comp Sci 2021;2.3:1–16.
[8] Zhang Jun. Deep learning based attack detection for cyber-physical system cybersecurity: A survey. IEEE/CAA J Automat Sinica 2021;9.3:377–91.
[9] Li Jian-hua. Cyber security meets artificial intelligence: a survey. Front Inform Tech Electron Eng 2018;19.12:1462–74.
[10] Ahsan Mostofa. Cybersecurity Threats and Their Mitigation Approaches Using Machine Learning—A Review. J Cybersecur Privac 2022;2.3:527–55.
[11] Zhao Shanshan. Computational intelligence enabled cybersecurity for the internet of things. IEEE Trans Emerg Topics Comput Intellig 2020;4.5:666–74.
[12] Zhang Zhimin. Artificial intelligence in cyber security: research advances, challenges, and opportunities. Artif Intellig Rev 2022;55.2:1029–53.
[13] Mostovoy YA, Berdnikov V. Large scale networks security strategy. Algorithms 2017;5.6:7–8.
[14] Stojanov Zeljko, Dobrilovic Dalibor. Exploring students’ experiences in using a physical laboratory for computer networks and data security. Comp Appl Eng
Edu 2017;25.2:290–303.
[15] Kumar Sunil, Agarwal Dilip. Hacking attacks, methods, techniques and their protection measures. Int J Adv Res Comp Sci Manage 2018;4.4:2253–7.
[16] Zhu Haibei. Operator strategy model development in UAV hacking detection. IEEE Trans Hum-Mach Syst 2019;49.6:540–9.
[17] Kokina Julia, Davenport Thomas H. The emergence of artificial intelligence: How automation is changing auditing. J Emerg Tech Account 2017;14.1:115–22.
[18] Salvagno Michele, Taccone Fabio Silvio, Gerli Alberto Giovanni. Can artificial intelligence help for scientific writing? Critic Care 2023;27.1:1–5.
[19] Talwar Rohit, Koury April. Artificial intelligence–the next frontier in IT security? Netw Secur 2017;2017(4):14–7.
[20] Anwar Amaan, Hassan Syed Imtiyaz. Applying artificial intelligence techniques to prevent cyber assaults. Int J Comput Intellig Res 2017;13.5:883–9.
[21] Maher Danny. Can artificial intelligence help in the war on cybercrime? Comp Fraud Secur 2017;2017(8):7–9.
[22] Jain Ankit Kumar, Sahoo Somya Ranjan, Kaubiyal Jyoti. Online social networks security and privacy: comprehensive review and analysis. Compl Intellig Syst
2021;7.5:2157–77.
[23] Zahra SR, Chishti MA. A generic and lightweight security mechanism for detecting malicious behavior in the uncertain Internet of Things using fuzzy logic- and
fog-based approach. Neural Comput Applic 2022;34:6927–52.
[24] Bhavsar M, Roy K, Kelly J, et al. Anomaly-based intrusion detection system for IoT application. Discov Internet Things 2023;3:5.
[25] Li Guangxu, Kou Gang, Peng Yi. Heterogeneous large-scale group decision making using fuzzy cluster analysis and its application to emergency response plan
selection. IEEE Trans Syst Man Cybernet Syst 2021;52.6:3391–403.

Qinghui liu was born in Jining Shandong, P.R. China, in 1977. he received the Master degree from Shandong University, P.R. China. Now, he works in Network Center
Zaozhuang University, His research interests include Information technology and big data analysis.

Tianping Zhang was born in Shanghai. China, in 1982. She works in School of Communication and Information Engineering, Shanghai Technical Institute of Electronics
& Information. Her research interests include artificial intelligence, and Internet of Things.

10