| dc.description.abstract |
In recent years, a new term has emerged which is now widely applied as IDS (Intrusion Detection System), namely Deep Learing. One type of Deep – Learing is RNN (Recurrent Neural Network) which has recently been applied to IDS. Cyber attacks cannot be avoided, but they can be anticipated by building a system than can detect the performance of network data flows so that users can avoid all kinds of attacks and intrusion attempts from unknown parties. This research aims to test and analyse the accurary and speed of the Recurrent Neural Network in detecting attacks. The method used for this research is RNN, which is operated through the Python and Google Colab programs. Based on the results, the model was trained with 50 epochs resulting in an accurary of 92%. Meanwhile, a model with 30 epochs produces an accurary of 99%. So, the model can work well on training data with a total of 30 epochs.
Keyword: recurrent neural network, intrusion detection system, network attacks
| en_US |
