基于击键压力和时间匹配的异常检测方法  

Anomaly detection based on matching pressure and time of keystroke

作　　者：李响[1] 朱参世[1] 任婷婷[1] 

机构地区：[1]空军工程大学工程学院航空装备管理系,陕西西安710038

出　　处：《计算机工程与设计》2010年第9期1935-1938,共4页Computer Engineering and Design

摘　　要：为了有效利用用户的生物特征进行入侵者身份识别,提出了一种基于用户击键特征进行异常检测的新方法。该方法根据人们在击键时所产生的按键压力和时间间隔的惟一性,利用正态分布的特性控制模式库生成方式,构造出能够描述每个用户独有特征的击键特征向量库,然后利用模式匹配算法对新登陆用户进行检测。相关实验验证了该方法具有较高的用户识别能力。To recognize the intruder effectively with biology characteristic, a new anomaly detection approach based on utilizing keystroke features is presented. The approach makes use of the unique of biologic characteristic which contains key pressure and keystroke frequency when users strike keyboard. The key pressure and keystroke frequency when user strikes the keyboard to enter his password are used to construct a peculiar keystroke vector representing each user respectively and controlled by the regulation of normal distributing. Pattern matching arithmetic is used to identify a new consumer. Finally, the approach is demonstrated to have a good discerning ability in the related experiment.

关 键 词：生物测定学 异常检测 模式匹配 击键压力 击键时间 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]