融合击键内容和击键行为的持续身份认证

Continuous authentication fusing keystroke behavior and keystroke content

为提高真实击键场景中用户的持续身份认证能力,搭建完全自由的实验环境采集击键数据。将连续击键事件中各后置击键的频次作为击键内容特征,将排序后的连续击键时间间隔序列作为击键行为特征,引入改进的Yager证据合成理论融合击键内容域和击键行为域的子分类器得到最终的持续身份认证模型。实验结果表明,与现有的击键认证模型相比,采用融合技术的认证方法提高了用户持续身份认证的准确率,在真实的内网中有应用价值。

To improve the user’s continuous identity authentication capability in real keystroke scenarios,a completely free experi-mental environment was built to collect keystroke data.The frequency of each post keystroke in the continuous keystroke event was taken as the keystroke content feature,and the sequence of consecutive keystroke time interval was taken as the keystroke behavior feature,and the improved Yager theory of evidence was introduced to fuse the sub-classifier of keystroke content domain and keystroke behavior domain to obtain the final continuous identity authentication model.Experimental results show that compared with the existing keystroke authentication model,the authentication method using fusion technology improves the accuracy of user’s continuous identity authentication,and it has application value in real intranet.