The computer security has become a major challenge. Tools and mechanisms\nhave been developed to ensure a level of compliance. These include the Intrusion\nDetection Systems (IDS). The principle of conventional IDS is to detect\nattempts to attack a network and to identify abnormal activities and behaviors.\nThe reasons, including the uncertainty in searching for types of attacks\nand the increasing complexity of advanced cyber-attacks, IDS calls for the need\nfor integration of methods such as Deep Neuron Networks (DNN) and Recurring\nNeuron Networks (RNN) more precisely long-term memory (LSTM).\nIn this submission, DNN and LSTM were used to predict attacks against the\nNetwork Intrusion Detection System (NIDS). In this memory, we used four\nhidden layers for all deep learning algorithms, forty-one layers of inputs and\ntwo layers of outputs and with 100 iterations. In fact, learning is kept constant\nat 0.01 while the other parameters are optimized. After that for DNN, the number\nof neurons of the first hidden layer was further increased to 1280 but did\nnot give any appreciable increase in accuracy. Therefore, the number of neurons\nhas been set to 1024 and the LSTM we set the number of neurons of all\nhidden layers to 32. The results were compared and concluded that a three-layer\nLSTM performs better than all other conventional machine learning and deep\nlearning algorithms.
Loading....