项目名称: 骨干网环境下基于访问行为及其无指导学习模型的网络失窃密异常检测
项目编号: No.61070185
项目类型: 面上项目
立项/批准年度: 2011
项目学科: 金属学与金属工艺
项目作者: 张永铮
作者单位: 中国科学院计算技术研究所
项目金额: 11万元
中文摘要: 骨干网环境下未知失窃密行为的发现具有重要的现实意义,然而,骨干网环境海量的数据特点给传统的异常检测方法带来了巨大的技术挑战,为此,本项目拟通过研究访问行为模型、行为属性无指导学习等关键模型与算法,解决无训练集的网络失窃密行为的异常检测问题,为进一步失窃密行为的追踪和控制提供关键理论与技术支撑。
中文关键词: 失窃密;访问行为;无指导学习;异常检测;骨干网
英文摘要: It is an important actual significance to detect unknown stealing or stolen secret behaviors under a backbone environment. However, the charactristic of mass data under the environment provides great challenge for traditional anomaly detection methods. Therefore, through studying access behavior model and unsupervised learning of behavior attributes, this project strives for solving the problem of anomaly detection for network stealing or stolen secret behavior without training sets, in order to provide the key theory and technology support for further tracking and controling.
英文关键词: stealing or stolen secret; access behavior; unsupervised learning; anomaly detection; backbone