NETWORK TRAFFIC ANALYSIS USING MACHINE LEARNING

Abstract

Network traffic analysis using Machine Learning (ML) has become an important research area in the field of cybersecurity and network management. With the rapid growth of internet usage, cloud computing, IoT devices, and online services, modern networks generate massive amounts of traffic data every second. Traditional traffic monitoring and rule-based detection systems are often unable to efficiently identify complex attack patterns, abnormal behavior, or evolving cyber threats. To overcome these limitations, machine learning techniques are widely used for intelligent network traffic analysis and anomaly detection. This project focuses on analyzing network traffic data using machine learning algorithms to detect malicious activities, classify traffic patterns, and improve network security performance. The system collects network traffic information such as packet size, protocol type, source and destination IP addresses, flow duration, and transmission rate. After preprocessing and feature extraction, the dataset is trained using supervised machine learning algorithms such as Decision Tree, Random Forest, Support Vector Machine (SVM), KNearest Neighbor (KNN), and Logistic Regression. These algorithms help in identifying normal and abnormal traffic behavior with high accuracy. The proposed system provides real-time traffic monitoring and threat detection by automatically learning from historical traffic data. It minimizes human intervention and improves the efficiency of intrusion detection systems. Machine learning models can recognize hidden patterns and detect unknown attacks that traditional systems may fail to identify. The system also enhances network reliability, reduces false alarm rates, and supports faster response to security incidents. Int. J. Engg. Res. & Sci. & Tech. 2026, ISSN 2319-5991 Vol. 22, No. 2, 2026 https://ijerst.org/index.php/ijerst 2887 The experimental results demonstrate that machine learning-based traffic analysis achieves better performance in terms of accuracy, precision, recall, and detection speed when compared to conventional methods. This approach can be effectively applied in enterprise networks, cloud environments, banking systems, educational institutions, and smart city infrastructures to ensure secure and reliable communication.