سامانه همایش های دانشگاه آزاد اسلامی واحد نجف آباد

صفحه اصلی / International Conference on Artificial Intelligence; City, Industry and Health

Network Anomaly Detection Using Artificial Intelligence Algorithms

نویسندگان :

Sajad Balali Dehkordi¹ Saeed Nasri² Sina Dami³

1- author 2- Corresponding author 3- 4Assistant Professor, Department of Midwifery, Nursing and Midwifery Sciences Development Resea

کلمات کلیدی :

Anomaly Detection،Bayesian Optimiziation،Intrusion Detection،Transformer algorithm،Network Traffic Flow

چکیده :

Nowadays, much attention has been devoted to the detection of intrusion and anomaly which is a primary step for the security of computer networks. Also, due to the high traffic volume in computer networks, anomaly detection has become a very important issue in recent years. The need for an effective anomaly detection system is significant for high-speed and accurate detection. Deep learning (DL) and machine learning (ML) methods are widely used in intrusion detection systems to detect and classify types of attacks. However, the high-dimensional data set in intrusion detection has reduced the speed and accuracy of deep and ML methods. The main challenge in designing intrusion and anomaly detection systems is attributed to the accuracy and training time of the deep and ML algorithms. The problem of speed in intrusion detection datasets can be solved using feature selection and feature reduction methods. As stated in the literature, the efficient design of computer-aided methods plays a prominent role in increasing the accuracy of abnormality detection for computer networks. Therefore, it is essential to provide an algorithm to enhance the detection of abnormality in the network traffic flow. Accordingly, the current paper aims to improve the detection of anomalies in network traffic employing the transformer algorithm and NSL-KDD dataset. The findings represent 95% accuracy achieved by the proposed method which is on par with its counterparts and even outperforms some of them like the CNNLSTM and EDA algorithms.