Packet Analysis for Malicious Activity Detection Using Machine Learning
Aryan Dinesh Deshmukh
Vishwakarma University, Pune, India
Abstrct:
This study investigates the detection of malicious and malformed network packets using supervised machine learning techniques. The research utilizes a subset of the UNSW-NB15 dataset to analyze network packet metadata. Key techniques such as k-nearest neighbors (kNN) and decision trees were implemented to differentiate between normal and malicious activities. The models achieved a classification accuracy exceeding 90%, highlighting their potential in enhancing cybersecurity defenses. This work underscores the necessity of continuous innovation in intrusion detection systems (IDS) to counteract evolving cyber threats.
Keywords:
Malicious packet detection, machine learning, k-nearest neighbors, decision trees, cybersecurity, network security
Published on: 06-2024
Journal Name: Science Management Design Journal
Volume: 02
Issue: 02
Pages: 56-64
Month: June
Year: 2024