面向用户行为的网络钓鱼智能防御研究

项目名称： 面向用户行为的网络钓鱼智能防御研究

项目编号： No.61304208

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 黄华军

作者单位： 中南林业科技大学

项目金额： 25万元

中文摘要： 钓鱼网站威胁已成为当前网络安全威胁的首位，网络钓鱼防御是其对抗技术。针对现有网络钓鱼防御存在性能、精度及实用性等不足，本课题提出面向用户行为的网络钓鱼智能防御研究，通过理解用户行为，当用户在浏览钓鱼网站时，激活轻量级检测引擎检测，采用提醒和混淆输入的防护机制。其研究内容包括:用户行为表征、建模和推理，钓鱼网站轻量级智能检测引擎和网络钓鱼智能处理机制.构建融合URL、交互性、网页噪声和站点Logo识别检测结果的轻量级检测引擎,解决钓鱼网站检测的性能与精度的不足;建立用户行为推理模型和智能处理机制,提高网络钓鱼防御工具的实用性.项目创新点包括:1)提出基于上下文感知的用户行为推理模型;2）提出基于权重的多重检测结果融合算法.本课题旨在用户行为理解、轻量级检测和智能处理方面完善网络钓鱼智能防御的理论与技术,建立及时、准确、自动检测钓鱼网站的智能防御新途径。

中文关键词： 网络钓鱼；网络钓鱼防御；用户行为；网页噪声；logo识别

英文摘要： Phishing attacks have been the top threat in Internet secuity. Anti-phishing is a countermeasure technology to phishing. Now a day, how to effectively combat phishing is a challenge in web security realm.To sovle the problems of performance, precision，practicability, this project proposese an intelligent anti-phishing based on user behavior.Firstly, the mechenicsm understands the user behavior. If the user browsers a phishing site, the lighting-weight engine will be actived. At last, a warning icon will be given and input information will be confused. The reasearch areas are as follows. 1) user behavior presentating, modeling and infering; 2) light-weight detection engine to detect phishing site; 3) intelligent process scheme to phishing. In this project, the detection results of URL, interactive, webpage noise, and websit logo recognition are fused to solve performance and precision shortages. To the practicability problem, the user behavior infering model and intelligent process scheme are used. Two new algorithms, including of a user behavior infering model based on context-aware, and a multi-detection results fusion algorithm based on weight, are proposed. This project aims at improving theory and technology to combat phishing, and providing a new "zero-hour" accuracy automatical intelligent anti-phish

英文关键词： phishing；anti-phishing；user behavior；webpage noise；logo recognition