基于改进免疫算法的BP神经网络在木马检测中的应用  

Application of BP Neural Network Model Based on Modified Immune Algorithm to Trojan Detection

作　　者：包晖[1] 

机构地区：[1]河南工业大学国际教育学院,郑州450001

出　　处：《现代计算机》2013年第15期3-5,共3页Modern Computer

基　　金：863计划(No.2012AA101608);河南省教育厅科技项目(No.13A520179)

摘　　要：在BP网络模型基础上,引入一种改进的免疫算法,采用并行计算方式,建立多特征区域匹配模式,每个特征区域产生一系列子检测器集合,经过迭代比较,产生成熟检测器集合,建立一个基于改进免疫算法的BP网络模型。将该模型应用于木马检测中,模型收敛速度快,响应时间短,检测率和误报率均优于传统算法。In allusion to the insufficiencies such as high missing report rate and false alarm rate, a kimt of BP neural network model is established based on the modified immune algorithm. Uses the par allel computing method and establishes a multi-feature matching pattern. Characteristics of each region produces a series of sub detector set. After iterative comparison, produces a collection of mature detector and eventually establishes a BP neural network model based on modified im mune algorithm. The model is applied to the Trojan detection, and it has a fast convergence speed and takes a short time to respose, and its＂ detection rate and false alarm rate are superior to the traditional algorithm＇s.

关 键 词：免疫算法 BP神经网络 木马检测 

分 类 号：TP309.5[自动化与计算机技术—计算机系统结构]