Computer Network Security Analysis Modeling Based on Deep Learning Algorithm

Abstract

Network security is an important factor affecting its popularity. After years of research and practice, many scholars and network security companies are committed to defense research and have achieved some success. However, firewalls, antivirus software and so on all have a certain degree of passivity, there is no active defense mode of real-time collection. In order to improve the performance of defense system, this paper introduces deep learning technology. Deep learning is a multi-level convolutional neural network, which can find potential and valuable data from massive data and apply it to network security defense. This can discover the virus or Trojan data in the network in time, so as to improve the initiative of network security defense. Based on the analysis of deep learning algorithm and the characteristics of intrusion detection data set, this paper proposes a normalized coding algorithm. The algorithm uses three hidden layer coding and a learning network to learn intrusion detection data. Three hidden layers realize the extraction and coding of input information, and the normalization layer realizes the normalization of results. Then, regression learning is carried out according to the comparison with the target results. Experimental results show that the algorithm can effectively improve the detection results.