微型电脑应用2025，Vol.41Issue(9)：80-83,4.
基于GAN-LSTM的APT攻防对抗过程滞后性分析模型
Analysis Model of APT Attack Defense Confrontation Process Hysteresis Based on GAN-LSTM
摘要
Abstract
In order to solve the hysteresis problem of current defense methods in responding to advanced persistent threat(APT)attacks,a generative adversarial network-long short-term memory(GAN-LSTM)-based analysis model for the hystere-sis of APT attack defense confrontation process is constructed.This paper obtains hysteresis analysis source data,uses infor-mation entropy as a metric,and selects hysteresis indicators from domain name system(DNS)logs.This paper utilizes GAN to generate adversarial samples,trains the LSTM analysis model to complete the construction of the hysteresis analysis model.The model is used to calculate the hysteresis index and determine the corresponding degree of hysteresis.The results indicate that the intersection and union ratio under the application of the constructed analysis model ranges from 0.8 to 1.0,which is the maximum value among the four methods,proving the accuracy of the constructed analysis model.关键词
生成对抗网络—长短期记忆/高级持续性威胁攻防对抗/滞后性描述指标/滞后性分析模型Key words
GAN-LSTM/APT attack defensive confrontation/hysteresis description index/hysteresis analysis model分类
信息技术与安全科学引用本文复制引用
滕子贻,周金磊,翟曲,孟勐..基于GAN-LSTM的APT攻防对抗过程滞后性分析模型[J].微型电脑应用,2025,41(9):80-83,4.基金项目
国网辽宁省电力有限公司科技项目资助(2022YF-101) （2022YF-101）
