基于协同进化神经网络的垃圾邮件识别研究
首发时间:2013-01-28
摘要:互联网上大量冲斥的垃圾邮件影响着企业和个人的正常工作与生活,垃圾邮件识别已成为信息安全的重要研究内容之一。本文针对垃圾邮件识别问题,提出了一种基于多目标协同进化的神经网络分类算法,首先利用K-means算法对传统训练方法确定的网络初始隐节点聚类,然后以聚类后的隐节点群生成子种群进行协同进化操作,采用多目标优化算法评价个体适应值,各子种群的代表个体共同组成网络的最优结构。算法采用包含整个网络隐节点结构和控制向量的矩阵式混合编码方式,神经网络权值由伪逆法求解确定。通过仿真实验验证,该算法快速有效,优于其他常用分类算法,具有较强的垃圾邮件识别能力。
For information in English, please click here
A Coevolutionary Neural Network Approach for Spam Classification
Abstract:The annoyance of spam increasingly plagues both individuals and organizations. Spam classification is an important issue in the field of the information security to distinguish the spam with the legitimate email. This paper presents an effective neural network approach based on a specially designed cooperative coevolution paradigm for spam classification. The K-means method is employed to divide the initial hidden nodes into modules that are represented as subpopulation of the coevolutionary algorithm. The multiobjective algorithm is adopted to evaluate the fitness of individuals. Collaborations among the modules are formed to obtain complete solutions. The algorithm adopts a matrix-form mixed encoding scheme with a control vector. The weights between the hidden layer and the output layer are calculated by pseudo-inverse algorithm. Experimental results illustrate that the proposed algorithm outperforms the traditional classification methods on the spam classification problems.
Keywords: RBFNN Coevolutionary algorithm Multiobjective optimality Spam classification
论文图表:
引用
No.****
同行评议
勘误表
基于协同进化神经网络的垃圾邮件识别研究
评论
全部评论0/1000