Intrusion Detection in Network Traffic Using LSTM and Deep Learning
Prabu Jayant, Mohit Shetty, Sannapaneni Jeevan, Mohana, Minal Moharir, Ankit Kumar · 2024
Network Intrusion Detection system (NIDS) helps in detecting the traffic and avoids any potential harm to the system. LSTM is a deep learning model architecture that uses recurrent neural networks (RNN) for handling sequential data. Unlike traditional IDS that depend on fixed signatures, LSTMs can interpret network traffic patterns and detect anomalies that indicate new attack methods. This flexibility makes it more effective against evolving cyber threats. While traditional IDS may only detect malicious activity, LSTMs can predict the specific type of attack by analyzing traffic sequences. This improved accuracy enables quicker and more targeted response strategies to prevent cyber threats. LSTMs have the capability to analyze network traffic data in real-time, allowing for immediate threat identification. The obtained accuracy and F1 score of the model are ${9 2. 8 3 \%}$ and ${9 4. 2 5 \%}$ respectively. This real-time analysis enables faster response times and classification of cyberattacks, reducing the potential impact of any security breaches.