This project focuses on detecting anomalies in network activity, specifically identifying Neptune (SYN flood) attacks using machine learning techniques.
The dataset contains detailed records of network activities, with each record labeled as either normal or a "Neptune" attack. This binary classification problem aims to distinguish between legitimate network traffic and potential security threats.
- Training set: 86,845 rows
- Test set: 21,712 rows
- Target variable: "Attack" (0 for normal, 1 for Neptune attack)
A Neptune attack, also known as a SYN flood attack, is a type of denial-of-service (DoS) attack. The attacker overwhelms a target system with a high number of SYN requests, exploiting the TCP handshake process to consume resources and render the system unresponsive to legitimate traffic.
This project uses Support Vector Machine (SVM) for classification. SVM is chosen for its effectiveness in high-dimensional spaces and its versatility through different kernel functions.