Anomaly Intrusion Detection System Using Neural Network
No Thumbnail Available
Date
2020
Authors
Raghad, Bashir
Hanan, Rayyan
Mohammad, Khateeb
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
security has always been a challenge for administrators, as bay by day attacks
became sophisticated. which make it harder for Intrusion detection systems (IDS) to detect,
there are three types of IDS, signature-based, anomaly-based, IDSs use signature-based and
anomaly-based is known as hybrid. Many researchers are focused on anomaly-based due to its
ability to detect zero-day-attack, but there are some limitations such as false positive alarms
problem. By using Arti cial Neural Networks (ANN), the process of designing and testing
AIDS is easier. In this paper, the Convolutional Neural Network (CNN) and Long Short Term
Memory (LSTM) neural network are used to reduce the number of false alarms, by using Keras,
sklearn API, and TensorFlow with CSE-CIC-IDS2017, 2018 datasets for training and testing the
AIDS model. In our research, we found that the accuracies of CNN and LSTM neural networks
are 93% and 99% respectively, and CNN is better in multi-classi cation while LSTM is better
in binary-classi cation.