基于生存簇识别和预测的生存态势感知模型

可生存系统安全态势的形势日趋严峻,增强生存性的前提是识别出目标系统的生存态势,该文构建了一种基于生存簇识别和预测的生存态势感知模型。首先,对生存态势数据采用Ward增强聚类法实现不同服务等级生存簇的分类和识别;其次,使用自回归积分滑动平均(ARIMA)模型预测目标系统生存态势的未来趋势,并对预测结果进行了残差修正;最后,结合事前识别和事后预测实现了对可生存系统生存态势的感知。仿真实验表明,该模型具有良好识别效果和较高的预测准确度。

面向智慧校园的学生异常行为检测

结合智慧校园数据挖掘中产生的空间位置信息,提出了一种基于消错决策和离群偏离度计算校园中学生行为异常检测方法,实验表明该方法不仅可以快速的识别学生单一时间点的异常行为,还可以实现对学生全天行为轨迹的异常判断,便于学校更好地保障学生的安全.

Modeling and analysis of cloud computing system survivability based on Bio-PEPA

For the cloud computing system,combined wth the memory function and incomplete matching of the biological immune system,a formal modeling and analysis method of the cloud computing system survivability is proposed by analyzing the survival situation of critical cloud services.First,on the basis of the SAIR(susceptible,active,infected,recovered)model,the SEIRS(susceptible,exposed,infected,recovered,susceptible)model and the vulnerability diffusion model of the distributed virtual system,the evolution state of the virus is divided into six types,and then the diffusion rules of the virus in the service domain of the cloud computing system and the propagation rules between service domains are analyzee.Finally,on the basis of Bio-PEPA(biological-performance evaluation process algebra),the formalized modeling of the survivability evolution of critical cloud services is made,and the SLIRAS(susceptible,latent,infected,recovered,antidotal,susceptible)model is obtained.Based on the stochastic simulation and the ODEs(ordinary differential equations)simulation of the Bio-PEPA model,the sensitivity parameters of the model are analyzed from three aspects,namely,the virus propagation speed of inter-domain,recovery ability and memory ability.The results showthat the proposed model has high approximate fitting degree to the actual cloud computing system,and it can well reflect the survivable change of the system.

基于空间态势感知的学生异常行为预警方法

大学生心理健康问题已经渐渐成为社会关注的焦点,本文以空间态势感知的行为轨迹分析技术为基础,设计出对具有心理疾病的人员进行及时定位的分析系统——迹忆,该系统利用基于空间态势的行为轨迹分析,运用离群偏离度的计算方法,能够实现对异常情况的及时预警,同时还能保护用户的隐私。