自然科学版
陕西师范大学学报(自然科学版)
专题研究
基于支撑矢量机和Windows Native API的异常检测方法
PDF下载 ()
余景景,强宁
(陕西师范大学 物理学与信息技术学院, 陕西 西安 710062)
余景景,女,博士研究生,研究方向为计算机网络管理及安全、智能信息处理.
摘要:
借鉴Unix类系统下基于系统调用的主机异常检测理论,通过追踪Windows本机应用编程接口调用序列,对Windows系统下的主机异常检测进行研究.在异常序列检测中,结合使用对小数据集具有较好推广能力的支撑矢量机方法,进而取得较高的检测准确率.实验表明Native API可为Windows平台下基于主机的异常检测系统提供一种可能的数据源.
关键词:
异常检测; 支撑矢量机; Windows 本机应用编程接口
收稿日期:
2007-05-05
中图分类号:
TP393.01
文献标识码:
A
文章编号:
1672-4291(2007)04-0037-04
基金项目:
Doi:
Native API Based Windows abnormal Detection Method Using SVM
YU Jing-jing, QIANG Ning
(College of Physics and Information Technology, Shaanxi Normal University, Xi′an 710062, Shaanxi, China)
Abstract:
According to the study of host abnormal detection based on system calls under UNIX-like systems, this paper completes the similar research via tracing the sequences of Windows Native APIs (Application Programming Interfaces, APIs) under Windows platform. In the process of abnormal sequence detection, the SVM(Support Vector Machine) method is used for its generalization capability in small-scale dataset and a high accuracy of detection is obtained. The experimental results show that Windows Native APIs are possible data source for the host abnormal detection system under Windows platform.
KeyWords:
abnormal detection; support vector machine (SVM); Windows native application programming interface (API)