Comparative Analysis of Neural Networks for Intrusion Detection System

1-3Dept. of Information Technology, Xavier Institute of Engineering, Mumbai, India 4Prof. Martina D’Souza Dept. of Information Technology, Xavier Institute of Engineering, Mumbai, India ---------------------------------------------------------------------***--------------------------------------------------------------------Abstract With rapid expansion in the range of computer hardware, networking and operating systems along with continuously changing capabilities and creativity of the attackers and ever-changing nature of threat to the systems, security of data has become a very crucial issue. Due to the exponential growth of network traffic data and modern attacks requirements, traditional Network Intrusion Detection Systems (NIDS) encounter difficulties. While Machine Learning algorithms show promising results, the Neural Networks have now gained popularity and are widely used for many applications. This paper presents a comparative analysis of four Neural Network based algorithms namelyCNN, DNN, LSTM and MLP to highlight the efficiency of the Neural Networks in detecting the attacks. For training and evaluating the intrusion, NSL-KDD dataset is used. The result of this comparative analysis has been found to be efficient with high accuracy and has a promising scope for further research.