Intrusion Detection System Using Deep Learning for DoS Attack Detection

Abstract

Various attacks on a computer network or the internet have generated many incidents and cases, this makes security threats in using the internet or computer networks a major focus. Denial of Service attack or often referred to as DoS attack is one of the attack techniques that carry out flooding packets or requests to the target computer until the target computer is down. Prevention is needed in order to minimize existing attacks. IDS can be used as a detector in network traffic, but because IDS has its limitations, an IDS system is built using Deep Learning to detect DoS attacks. By using the data from the wireshark log as a dataset, it is necessary to do data normalization which will then be inputted into CNN VGG-19. The test results that have been carried out with variations in the data inputted into the CNN VGG- 19 produce an average accuracy of 99.32% with an average loss of 4.08%, and by varying the iteration of the training process the resulting accuracy is 99.17% with an average loss - an average of 4.46%. And the ROC Curve value for the True Positive Rate and the False Positive Rate is 1.