International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
Network Intrusion Detection System using Machine Learning Techniques K. Neeraj Reddy#1, S. Harish Reddy#2, P. Vamshi#3, A.Ashwini #4 # Department of CSE, Vardhaman College of Engineering, Telangana, India. ---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - The constant progress of technology has
computer systems are combined into an intrusion detection system (IDS). It is an IDS that may be used as a hybrid, exception-based, or signature system. An intrusion detection system (IDS) that uses signatures compares proposed properties with pre- established intrusion patterns to find intrusions. In contrast, anomaly-based intrusion detection systems concentrate on proficiently interpreting actions to detect deviations. A variety of methods (e.g., statistics-based, data- based, and artificial intelligence-based, including recent research on deep learning) are used to detect implausibility.
brought forth a lot of advantages and major difficulties for several businesses, chief among being cybersecurity. Robust intrusion detection systems (IDS) are necessary to protect against hostile activities due to the increase in cyber threats. In this study, we identify potential intrusions using machine learning techniques, namely the Support Vector Machine (SVM) algorithm, using the CICIDS2017 dataset. Our testing produced encouraging results, with SVM recognizing pilot port incursions with an accuracy of 97.80%. We also investigate the effectiveness of other algorithms, including Random Forest, Convolutional Neural Networks (CNN), and Artificial Neural Networks (ANN), which shown different accuracy levels between 63.52% and 99.93%. Our proposed system encompasses data collection, preprocessing, training, and testing modelling, culminating in the development of an attack detection model. The system holds immense potential in fortifying networks against malicious attacks, removing or securing malicious content, and ensuring the confidentiality of sensitive information.
The field of malpractice in computers has evolved and continues to move beyond trivial behaviour, such as monitoring access to information, primarily to uncover larger threats. Data protection from unauthorized use, disclosure, alteration, destruction, and corruption is known as information security. The words "data security," "computer security," and "data protection" are sometimes used synonymously. These interactive lists are made to guarantee information availability, confidentiality, and integrity. Empirical studies indicate that gathering and identifying system information is the initial stage of an attack. Checking open ports in the system is important information for attackers and requires the use of various tools such as antivirus software and IDS. Recently, an IDS model for port testing that captures knowledge and tactics has been developed using machine learning and support vector machine (SVM) methods.
Key Words: Machine Learning, KDD, Cyber Security, Network, SVM, Random Forest.
1.INTRODUCTION Recent years have seen significant changes in many areas of connected technology, including smart grids, the Internet of Things (IoT), long-term development and 5G communications. The number of devices connected to an IP network is expected to exceed ten times the world's population by 2022, generating 4.8 ZB of IP traffic per year. Because a lot of sensitive information is transported across the unreliable "Internet" utilizing dated and inconsistent technology and communication protocols, this fast expansion raises severe security problems. To ensure the safety and security of the Internet, higher security measures and potential analysis should be carried out in the early stages of deployment.
2. RELATED WORK This article provides an overview of current developments in the subject, with a particular emphasis on studies on the use of NSL-KDD data. Therefore, in order to promote improved comparability of research in the literature, all aspects stated subsequently should be handled in accordance with NSL-KDD. The majority of the research employed training and testing data, which is a significant restriction. Lastly, several search algorithms based on deep learning for comparable problems are explored. One of the earliest studies was the development of an artificial neural network (ANN) to identify intrusion detection systems (IDS) with enhanced recovery capabilities.
Attacks must be prevented, detected, and responded to by the security mechanisms that have been put in place. intrusion. detection systems (IDS) are extensively employed to detect and identify any malicious activities and internal and external network attacks, along with any irregularities that may point to an incursion. Tools and methods for keeping an eye on network traffic and
© 2024, IRJET
|
Impact Factor value: 8.226
This approach uses training data only for training (70%), validation (15%), and testing (15%); this results in decreased performance when using unnamed for testing.
|
ISO 9001:2008 Certified Journal
|
Page 642