首页|基于高维多目标序贯三支决策的恶意代码检测模型

基于高维多目标序贯三支决策的恶意代码检测模型

扫码查看
原文链接
  • 万方数据
  • 维普
针对传统基于二支决策的恶意代码检测方法在面对动态环境中的复杂海量数据时,没有考虑在信息不充足条件下进行决策产生影响的问题,本文提出了一种基于卷积神经网络的序贯三支决策恶意代码检测模型.通过卷积神经网络对样本数据进行特征提取并构建多粒度特征集,引入序贯三支决策理论对恶意代码进行检测.为改善检测模型整体性能,避免阈值选取的主观性,本文在上述模型的基础上,同时考虑模型的综合分类性能、决策效率和决策风险代价建立高维多目标序贯三支决策模型,并采用高维多目标优化算法对模型进行求解.仿真结果表明,模型在保证检测性能的同时,有效地提升了决策效率,降低了决策时产生风险代价,更好地拟合了真实动态检测环境.
Malicious code detection model based on high-dimensional multi-objective sequential three-way decision
In view of the problem that traditional two-way decision based malicious code detection methods fail to con-sider the impact of decision making under the condition of insufficient information when facing complex and massive data in a dynamic environment,this paper proposes a sequential three-way decision malware detection model based on convolutional neural network.Firstly,the features of sample data were extracted and multi-granularity feature sets were constructed through convolutional neural networks,and then the sequential three-way decision theory was introduced to detect malicious code.To improve the overall performance of the detection model and eliminate the subjectivity of threshold selection,a high-dimensional multi-objective sequential three-way decision model was built based on the above model,taking account of the comprehensive classification performance,decision efficiency and decision risk cost of the model.In addition,the high-dimensional multi-objective optimization algorithm was used to solve the model.The simulation results show that the model can not only guarantee the detection performance,but also effectively improve the decision efficiency and reduce the decision risk cost.It better fits the real dynamic detection environment.

malware detectionsequential three-way decisionconvolutional neural networkhigh-dimensional multi-objective optimizationNSGA-IIImulti-granularitydelay decisiondecision threshold

崔志华、兰卓璇、张景波、张文生

展开 >

太原科技大学 大数据分析与并行计算山西省重点实验室, 山西 太原 030024

中国科学院 自动化研究所,北京 100089

恶意代码检测 序贯三支决策 卷积神经网络 高维多目标优化 基于参考点的高维多目标进化算法 多粒度 延迟决策 决策阈值

国家自然科学基金中央财政指导地方科技发展基金中国高校产学研创新基金-未来网络创新研究与应用项目太原科技大学研究生联合培养示范基地项目

61806138YDZJSX2021A0382021FNA04014JD2022003

2024

10.11992/tis.202306013

智能系统学报
中国人工智能学会　哈尔滨工程大学

智能系统学报

CSTPCD北大核心
影响因子：0.672
ISSN：1673-4785
年,卷(期)：2024.19(1)
  • 27