基于普适环境下陌生交互实体之间没有先验交互经历,为了该环境的计算安全,必须确保其交互的行为可信,提出一种基于嵌套的狄利克雷过程和无限隐马尔可夫模型的行为分析与态势预测模型。该模型能有效地预防陌生实体交互过程中反常、欺诈等行为,对不可信的行为提前干预,避免产生严重的后果。最后举出模型在智能商场中应用并仿真分析了模型的可行性。
There is no priori interactive experience between the strange interactive entities in the pervasive environment, so must make sure that the interactive behavior between the two entities is reliable for the computing safety in this environment. Based on this, this paper proposed a behavior-analyzing and situation-predicting model based on the nested Difichlet process and the infinite hidden Markov model. It could predict the behavior of abnormality and cheat effectively in the entity' s strange interactive process and intervene the unreliable behaviors in advance so that serious consequences could be avoided. Took some examples about the model' s application in the intelligent market and analysed the feasibility of model by simulated.