项目名称: 在线社交网络上恶意网址的实时预警
项目编号: No.61472162
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 自动化技术、计算机技术
项目作者: 李强
作者单位: 吉林大学
项目金额: 81万元
中文摘要: 在线社交网络(OSN)迅速增长的用户群和信息量,成为攻击者的新目标。OSN帖子中的恶意网址(URL)可用于发起多种攻击,如垃圾邮件、钓鱼、传播恶意软件等,且速度更快、范围更广,造成的危害更为严重。然而,由于攻击者利用OSN的特性和使用短网址等更复杂的攻击手段,现有恶意URL检测方法的准确率尚待提高,实时性不够好,对攻击技术变化的适应性也不够强,并且适用的OSN都是国外的,国内用户无法利用。本项目提出OSN中恶意URL实时预警技术的研究课题,充分利用OSN的特点改进恶意URL预报的准确率,使用改进的机器学习方法和客户-服务员系统提高预报速度,采用动态检测技术,对恶意URL库、训练样本和分类模型进行自适应变化,及时更新。要在用户点击恶意URL之前得到预警,最大限度减少漏报率和误报率。最终目标是提出一个适用于我国OSN上恶意URL实时预警应用软件,及时保护用户安全和遏制恶意URL的传播。
中文关键词: 在线社交网络;恶意网址;实时预警
英文摘要: With the rapid growth of users and information on the online social networks (OSN), it has become the new target of attackers. Malicious URLs in OSN's posts can be used to launch various attacks such as spamming, phishing, propagating malwares, etc . These attacks are faster, broader and cause more serious damages. However, because of OSN's special features and more complex means that attackers have exploited, such as the shorten URLs, there are many challenges in current malicious URL detection approaches, such as higher overheads, lower detection accuracy and realtime perfomance, not easy to detect the changes of attack methods. Especially, these approaches aim at foreign OSNs thus we are not able to utilize these approaches in Chinese OSNs. This project proposes prewarning malicious URLs in realtime on an OSN. It will take full advantages of Chinese OSN's features to improve the accuracy of prewarning malicious URLs. It uses improved machine learning algorithms and a sever-client system to improve the prewarning speed. It uses dynamic detection approaches in adaptive update of malicious URLs database, training samples and classification models. It can prewarn before a user clicking the URL and reduces the false positive rate and false negative rate as much as possible. The final goal is to propose and develop an application for prewarning malicious URLs in realtime on an OSN suitable for our country to protect users' security and block malicious URLs' propagation.
英文关键词: Online Social Network(OSN);Malicious URLs;Prewarning in realtime