Deep Learning-Based Intrusion Detection System (IDS) for Detecting Cyber Attacks in Network Traffic

Abstract

Network security has become a significant concern in the modern digital era due to the rising Complexity and the frequency of cyber-attacks now days.. Nonetheless, the main function of intrusion detection systems (IDS) is to keep monitoring on network traffic and spot malicious activities. The research examines a hybrid strategy to IDS by combining traditional machine learning techniques such as Support Vector Machines and Random Forest, along with the K-Nearest Neighbors, with deep learning models, specifically Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. The research relies on the CIC-IDS2017 dataset as their source , which consists of an extensive collection of attack and normal traffic records, to assess the performance of the models are been working A variety of metrics are used to evaluate the implemented to the models’ performance, including recall, precision, accuracy, and the F1 score. The results of the analysis indicate that deep learning models, especially LSTM, have a better performance than traditional machine learning methods. This result implies the main advantage of the deep learning being ability to really capture the complex patterns in the network traffic thus providing a stronger and reliable intrusion detection approach. The research contributes to the developments of intelligent IDS frameworks that can enhance cyber security in environments for real-time networks