The increasing sophistication of cyberattacks has rendered traditional intrusion detection systems inadequate for safeguarding modern network infrastructures. This research presents a real-time Network Intrusion Detection System (NIDS) leveraging Artificial Neural Networks (ANNs) to improve detection accuracy for both known and emerging threats. Unlike traditional models such as Support Vector Machines (SVM) and Random Forest, which rely on predefined rules, our ANN model dynamically learns attack patterns, making it more adaptable to novel threats. The model was trained using the CICIDS 2018 dataset, which includes various attack types such as Distributed Denial of Service (DDoS), Phishing, SQL Injection, and Brute Force attacks. To ensure real-time performance, the system was implemented using TensorFlow.js, enabling efficient in-browser threat detection without relying on backend processing. Performance evaluations indicate that the model achieved an accuracy of 90%, surpassing conventional techniques in detecting complex attack patterns. Additionally, real-time deployment with React.js and TensorFlow.js allowed seamless visualization and threat analysis, demonstrating scalability and efficiency. However, higher false positive rates were observed in Phishing detection, suggesting the need for future improvements, such as enhanced feature extraction and ensemble learning techniques. This study contributes to the evolution of intrusion detection systems by offering a scalable, adaptive, and real-time cybersecurity solution capable of mitigating emerging cyber threats.

Top