期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2016
卷号：9
期号：4
页码：41-50
DOI：10.14257/ijhit.2016.9.4.04
出版社：SERSC
摘要：Context information in smart space are complex and uncertain, it increases the difficulty of information prediction. In order to increase the accuracy of information prediction, a new method is proposed in this paper. Firstly, first-order predicate logic is used to describe the context information and Markov logic network is used to achieve the unification of the first-order logic and probabilistic graphical models. Secondly Markov logic network model is established based on some particular scenes which were predicted by neural network, the purpose of this model is to realize the prediction of device information which is valuable for the task planning by relevant complex context information. At last, simulation test is done using the real data collected from our smart space laboratory, the experiment results show that method is effective.
