e-ISSN: 2582-5208

International Research Journal of Modernization in Engineering Technology and Science

( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:11/November-2023 Impact Factor- 7.868 www.irjmets.com

DETECTION OF CYBER ATTACKS AND NETWORK ATTACKS USING

MACHINE LEARNING ALGORITHMS
Shradha Farane*1, Rutuja Gotane*2, Sakshi Chandanshive*3, Srushti Khandekar*4,
Khushi Agrawal*5
*1Professor, Department Of Information Technology, Jspm’s Jayawantrao Sawant College Of
Engineering, Pune, Maharashtra, India.
*2,3,4,5Student, Department Of Information Technology, Jspm’s Jayawantrao Sawant College Of
Engineering, Pune, Maharashtra, India.
ABSTRACT
The Internet and computer networks have become an important part of organizations and everyday life. New
threats and challenges have emerged to wireless communication systems especially in cyber security and
network attacks. The network traffic should be monitored and analyzed to detect malicious activities and
attacks to ensure reliable functionality of the networks and security of users’ information. Machine learning
techniques can be applied to detect the network attacks. In cyber security, intrusion detection, spam detection,
malware classification and detection, and phishing detection can be performed using machine learning
approaches. Although ML cannot fully automate a cyber-security system, it can identify cybersecurity threats
more efficiently than other software-oriented approaches, relieving security analysts of their burden. As a
result, machine learning techniques, can yield higher detection rates, lower false alarm rates, and cheaper
computing and transmission costs. Our goal is detection of cyber security and network attacks such as IDS,
phishing and XSS, SQL injection. The proposed procedure in this study is to structure the deep neural networks
for detection, which should detect the attack's existence in the early stages of the attack.
Keywords: Cyber Security, Network Attacks, Phishing, XSS, SQL Injection, IDS.
I. INTRODUCTION
In this modern era of information and communication technologies, physical objects are now connected with
each other through cyber networks are collectively called cyber physical system. Attack detection and
prevention, simply detects and prevents attacks in network traffic. An exploit is a tool or a software code or
method that takes advantage of a vulnerability to carry out malicious tasks.
Nowadays internet and computer networks play a crucial role in organizations and in everyday life. With the
increase in our dependence on computers and communication networks, malicious activities have become
increasingly prevalent. Network attack have become a vital problem in present communication environment.
The network traffic should be monitored and analyzed to detect malicious activities and attacks to ensure
reliability of the network and security of users’ information. Nowadays, machine learning techniques have been
applied to detect network attacks. Machine learning models are able to extract similarities and patterns in the
network traffic, there is no need to analyze manually for extracting attack patterns. Machine learning
algorithms can automatically build predictive models to detect network attacks.
In this system, we offer a review on attack detection methods involving strength of deep learning techniques.
First of all we summarize fundamental problems of network security and attack detection and then introduce
several related applications using deep learning. On the basis of categorization on deep learning methods, we
pay special attention to attack detection methods built on different kinds of architectures, such as auto-
encoders, generative adversarial network, recurrent neural network, and convolutional neural network. Later,
we present some datasets with descriptions and compare the performance to represent the approaches to
show current state of working of attack detection methods with deep learning. Finally, we summarize this work
and discuss some ways to improve the performance of attack detection under thoughts of utilizing deep
learning structures.
II. METHODOLOGY
The methodology for this project involves following key steps:

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[3357]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:11/November-2023 Impact Factor- 7.868 www.irjmets.com
Data Collection:
Collect diverse and accurate dataset for XSS, SQL Injection, Phishing attack and IDS. Ensuring the datasets
represent diverse attack scenarios and network traffic patterns.
Data Preprocessing:
Clean and preprocess the raw data, which may include network traffic logs, system logs, or packet captures.
Feature Selection and Feature Extraction:
Involves choosing a subset of the most relevant features (variables) from the original dataset. Feature
extraction involves transforming the original data into a lower-dimensional space while retaining as much
relevant information as possible.
Algorithm Selection:
Choose suitable machine learning algorithms based on the characteristics of the dataset and the nature of the
attacks.
Result Prediction:
Goal of result prediction is to accurately determine whether a given data instance corresponds to a legitimate
network behavior or an attack.
III. MODELING AND ANALYSIS
Cybercrime is spreading throughout the world, using any sort of weak spot in the computing environment.
ethical hackers are usually involved with assessing vulnerabilities and offering mitigation techniques. The
improvement of effective techniques is a pressing need in the cybersecurity community. maximum techniques
used in nowadays IDS aren't capable of addressing the dynamic and complicated nature of cyber-assaults on
laptop networks. gadget studying for cyber safety has emerged as a trouble of first-rate significance recently
because of the effectiveness of system learning in cyber safety issues. ML tactics have been used to deal with
critical problems in cyber security, including intrusion detection, malware type and detection, junk mail
detection, and phishing detection. even though machines gaining knowledge cannot fully automate a cyber-
security device, it can identify cyber safety threats with greater effects than other software-orientated methods,
easing the pressure on security analysts. As a result, effective adaptive procedures, which include machine-
gaining knowledge of strategies, can bring about higher detection costs, decrease fake alarm charges, and
cheaper computing and transmission prices. Our principal intention is to show that the trouble of detecting
assaults is fundamentally one of a kind from those other programs, making it far extra hard for the intrusion
detection network to properly use machine-gaining knowledge of gadgets gaining knowledge of algorithms
may be used to teach and hit upon if there was a cyber assault. As quickly as the assault is detected, an e-mail
notification can be dispatched to the safety customers.

Figure 1: System Architecture

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[3358]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:11/November-2023 Impact Factor- 7.868 www.irjmets.com
Any category set of rules may be used to decide whether or not or not an attack is a DoS/DDoS assault. assist
Vector system (SVM), a supervised gaining knowledge of method that analyses information and recognizes
styles, is one example of a category set of rules. for the reason that we cannot expect whilst, when, or how an
assault will occur, and absolute prevention cannot be assured, our excellent guess in the meantime is early
discovery, which allows you to assist reduce the danger of irreparable harm such occurrences can do.
corporations can use current answers or build their very own to come across cyber-assaults at a very early
stage to minimize the effect. Any gadget that calls for minimum human intervention could be ideal.
IV. RESULTS AND DISCUSSION
For IDS if the output is an anomaly, then it will be considered an attack, on the other hand, if the output is
normal then it is a legitimate packet.
For SQL Injection, Phishing attacks, and Cross Site Scripting attacks the output is in the format of 0 and 1, where
0 is not an attack and 1 will be considered malicious
Table 1. Comparative Analysis
SN. Attack Algorithm Accuracy
1 Intrusion Detection Decision Tree 99.47%
KNN Classifier 99.16%
BNB Classifier 90.67%
2 SQL Injection Attack Logistic Regression 92.85%
3 Cross Site Scripting Attack Convolutional Neural Network 98.59%
4 Phishing Attack Support Vector Machine 82.63%

Figure 2: Accuracy Score of various algorithms

V. CONCLUSION
In this study, an attempt was made to use the resilient control consensus method in complex discrete cyber-
physical networks with several local attacks off. By applying this control method, it was observed that even in
the presence of cyber-attacks, the system can remain stable and isolate the attacked node, and the performance
of the system is not weakened. Using the neural network used in this study, it was observed that with a deep
neural network, with 7 hidden layers, the system shows better performance. Also, in a recurrent neural
network integrated with a deep neural network, a deep layer network with a linear function performs better.
So, with the deep learning method, systems can analyze patterns and learn from them to help prevent similar
attacks and respond to changing behavior. To summarize, ML has the potential to make cyber security simpler,
more proactive, less expensive, and considerably more successful. After observing the state of the system
reported by the neural network, the control system makes decisions based on it and, if there is an attack,
www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science
[3359]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:11/November-2023 Impact Factor- 7.868 www.irjmets.com
detects it. As a result, effective adaptive approaches, such as machine learning techniques, can result in higher
detection rates, lower false alarm rates, and cheaper computing and transmission costs. We reviewed various
algorithms for attack detection based on various ML techniques. Because of the characteristics of ML
approaches, it is feasible to construct attacks with high detection rates and low false positive rates, while the
system rapidly adapts to changing hostile behaviors. One thing is sure, any organization failing to adopt these
techniques now or in the immediate future risk compromising data or worse servers.
ACKNOWLEDGEMENTS
We would prefer to give thanks to the researchers likewise publishers for creating their resources available. We
are conjointly grateful to the guide, and reviewer for their valuable suggestions and also thank the college
authorities for providing the required infrastructure and support.
VI. REFERENCES
[1] Z. N. Zarandi and I. Sharifi, "Detection and Identification of Cyber-Attacks in Cyber-Physical Systems
Based on Machine Learning Methods," 2020 11th International Conference on Information and
Knowledge Technology (IKT), 2020, pp. 107-112, doi: 10.1109/IKT51791.2020.9345627.
[2] Nurjahan, F. Nizam, S. Chaki, S. Al Mamun and M. S. Kaiser, "Attack detection and prevention in the
Cyber-Physical System," 2016 International Conference on Computer Communication and Informatics
(ICCCI), 2016, pp. 1-6, doi: 10.1109/ICCCI.2016.7480022.
[3] Ding Chen, Qiseng Yan, Chunwang Wu, and Jun Zhao, “SQL Injection Attack Detection and Prevention
Techniques Using Deep Learning,” Journal of Physics: Conference Series, Volume 1757, International
Conference on Computer Data and Artificial Intelligence (ICCBDAI 2020) October 2020, Changsha,
ChinaGyusoo Kim and Seulgi Lee, “2014 Payment Research”, Bank of Korea, Vol. 2015, No. 1, Jan. 2015.
[4] Ercan NurcanYılmaz, SerkanGönen, “Attack detection/prevention system against cyber-attack in
industrial control systems,” Computers & Security Volume 77, August 2018, pp 94-105
[5] Arpitha. B, Sharan. R, Brunda. B. M, Indrakumar. D. M, Ramesh. B. E, “Cyber Attack Detection and
notifying system using ML Techniques,” International Journal of Engineering Science and Computing
(IJESC), Volume 11, Issue No.06
[6] Yirui Wu, Dabao Wei, and Jun Feng, “Network Attacks Detection Methods Based on Deep Learning
Techniques: A Survey,” Security Threats to Artificial Intelligence-Driven Wireless Communication
Systems, 2020.
[7] Rafał Kozik, Michał Choraś, “Machine Learning Techniques for Cyber Attacks Detection,” Image
Processing and Communications Challenges 5, pp 391-398, Springer International Publishing
Switzerland 2014.
[8] Nutjahan, Farhana Nizam, Shudarshon Chaki, Shamim Al Mamun, M. Shamim, “Attack Detection and
Prevention in the Cyber Physical System,” 2016 International Conference on Computer Communication
and Informatics (IEEE -2016), Jan. 07 - 09, 2016, Coimbatore, India
[9] Anupong, W., Yi-Chia, L., Jagdish, M., Kumar, R., Selvam, P. D., Saravanakumar, R., & Dhabliya, D. (2022).
Hybrid distributed energy sources providing climate security to the agriculture environment and
enhancing the yield. Sustainable Energy Technologies and Assessments, 52
doi:10.1016/j.seta.2022.102142

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[3360]