Real-time Network Attack Detection Method based on Neural Network

The traditional convolutional neural network needs more advantages in dealing with real-time network intrusion. This paper proposes a new detection method using a visualization method when real-time traffic data is missing. We built a real-time traffic detection platform and invented a new algorithm structure: The system uses offline data to digitize and pixelate the hacked network traffic and then carries out gray-scale mapping, a pre-processing step. The data is rotated to the right and connected to the CNN interface for training. We use real-time network data to verify the experiment and evaluate our proposed model. The results show that the scheme proposed in this paper has a good detection effect and can detect real-time network intrusion types.

I. Introduction

Deep learning algorithms can learn the features of data samples adaptively under the support of mathematical principles. Convolutional Neural Network (CNN) has become a research hotspot in image processing, natural language processing, and other fields by its superior ability to extract features through autonomous learning, enabling various applications to achieve a leap in accuracy. Especially in network attack and recognition, it combines traditional algorithms’ feature extraction and classification process. It can achieve arbitrary complex nonlinear expression through model self-learning, avoiding the disadvantages of manual feature extraction, solving the modeling difficulties faced by traditional target recognition, and helping to improve the quality of the field of network attack recognition [1–2]. Therefore, deep learning algorithms have also been widely studied and applied in network attack identification.