摘要
分析用户网络行为,识别诈骗网站,对电信网络诈骗犯罪的防控工作具有重要的实际意义。本文通过研究用户的多步骤、时序性的网络数据流,提取显性流量特征以及隐性网络行为特征,建立基于BP神经网络的诈骗网站识别模型,实现对诈骗网站的动态识别。为解决传统的诈骗网站检测模型精确度不足、模型开销庞大的问题,以及传统BP神经网络模型容易陷入局部最优解的问题,本文提出遗传算法优化的BP神经网络模型,通过遗传算法寻找最佳初始权阈值,优化神经网络模型。在用户网络行为数据集的基础上进行实验,证明模型能够有效识别网站的诈骗属性,具有实战意义。
Based on the analysis of network behavior fraud website identification technology research,analysis of user network behavior,identify fraud website,has important practical significance for the prevention and control of telecom network fraud crime.In this paper,by studying the multi-step and sequential network data flow of users,we extract the characteristics of explicit traffic and im⁃plicit network behavior,and establish a fraud website identification model based on BP neural network to realize the dynamic identifica⁃tion of fraud websites.In order to solve the problems of insufficient accuracy and huge cost of traditional fraud website detection model,as well as the problem that traditional BP neural network model is easy to fall into the local optimal solution,this paper proposes a BP neural network model optimized by genetic algorithm,which uses genetic algorithm to find the best initial weight threshold and opti⁃mize the neural network model.Experiments on the basis of user network behavior data set show that the model can effectively identify the fraud attribute of the website,and has practical significance.
作者
连远博
盛蒙蒙
袁莹
周胜利
LianYuanbo;Sheng Mengmeng;Yuan Ying;Zhou Shengli(Department of Computer and Information Security,Zhejiang Police College,Hangzhou 310000)
出处
《现代计算机》
2021年第28期33-38,44,共7页
Modern Computer
关键词
网络行为分析
遗传算法优化
诈骗网站识别
internet behavior analysis
genetic algorithm optimization
fraud web application identification
