Cyber Attack Prediction in Enterprise Networks Using Temporal Convolutional Networks (TCN)

Recent cyberattacks on enterprise networks are increasingly frequent and sophisticated, threatening data privacy and business continuity. Detecting network attacks before they occur is crucial for enhancing network security. Traditional models, such as RNNs and LSTMs, struggle with gradient reduction and adapting to large time-based datasets. This study introduces a Temporal Convolutional Network (TCN) framework, which analyses long-term time patterns through specialised convolutions. The TCN design evaluates network traffic data to identify unique attack signals, enabling proactive defences as an enterprise intrusion detection system. Test results indicate that TCN outperforms other methods in accuracy, training speed, and adaptability. Overall, TCNs have proven to be a robust solution for detecting cyberattacks in dynamic enterprise environments.

Перевод пока недоступен