信息安全研究 ›› 2019, Vol. 5 ›› Issue (6): 495-499.
• 学术论文 • 上一篇 下一篇
王辉1,2,周忠锦1,2,王世晋1,2,史卓颖1,2
收稿日期:
出版日期:
发布日期:
通讯作者:
作者简介:
Received:
Online:
Published:
摘要： 传统的DGA攻击检测方法已经无法满足对不断变种的DGA域名的识别，检出准确率较低.因此主要研究一种基于MLP深度学习算法的DGA准确识别技术，通过已有的DGA样本数据集，提取多维度的特征向量信息，通过归一化、降维处理后，将特征向量输入MLP多层感知器进行训练，MLP多层感知器主要由输入层、隐藏层和输出层组成，训练后生成模型文件即可载入用于判断待检测的域名是否为DGA域名，可以有效提升DGA检测识别的准确度.
关键词: 域名生成算法(DGA), 多层感知器(MLP), CC服务器, 隐藏层, 奇异值分解算法
Abstract: The traditional DGA attack detection method can not meet the recognition of the everchanging DGA domain name, and the detection accuracy is low. This paper mainly studies an accurate DGA recognition technology based on MLP deep learning algorithm. Through the existing DGA sample dataset, multiple feature vector information is extracted. After normalization and dimensional reduction processing, the feature vector is inputted into the MLP, the MLP mainly consists of an input layer, a hidden layer and an input layer. The model file can be generated after training to determine whether the domain name to be detected is a DGA domain name, which can effectively improve the accuracy of the DGA detection and recognition.
Key words: domain name generation algorithm (DGA), multi-layer perceptron (MLP), command and control server, hidden layer, singular value decomposition
王辉 周忠锦 王世晋 史卓颖. 基于MLP深度学习算法的DGA准确 识别技术研究[J]. 信息安全研究, 2019, 5(6): 495-499.
0 / / 推荐
导出引用管理器 EndNote|Ris|BibTeX
链接本文: http://www.sicris.cn/CN/
http://www.sicris.cn/CN/Y2019/V5/I6/495
