Search

The online anomaly detection has been propounded as the key idea of monitoring fault of large-scale sensor nodes in Internet of Things. Now, the exciting progresses of research have been made in online anomaly detection area. However, the highly dynamic distributing character of Internet of Things makes the anomaly detection scheme difficult to be used in online manner. This paper presents a new online anomaly learning and detection mechanism for large-scale service of Internet of Thing. Firstly, our model uses the reversible-jump MCMC learning to online learn anomaly-free of dynamics network and service data. Next, we perform a structural analysis of IoT-based service topology by network utility maximization theory. The results of experiment demonstrate the method accuracy in forecasting dynamics network and service structures from synthetic data. [ABSTRACT FROM AUTHOR]

The coordinated learning is importance of technique for cooperative multi-objects system in large-scale Internet of Things . The coordinated learning has attracted a lot of attention for its applications in Internet of Things. However, the self-adaptive makes the coordinated learning difficult to be used in IoT. This paper proposes multi-objects scalable coordinated learning algorithm based on the maximum potential loss of coordination. The algorithm defines an interaction measure that allows objects to dynamically estimate the potential utility loss of coordination with any cluster of objects. The interaction mechanism makes each object compute their beneficial coordination set in different situations and makes the best use of their limited communication resource in Internet of Things. As a result of experiments, our algorithm adapts policy learning of object and their coordination network for different context. Finally, the experiments with the smart agriculture data set demonstrate that the proposed scheme is effective and robust. [ABSTRACT FROM AUTHOR]