The Research Of Technologies For Uncertainty Complex Event Process Based On Markov Logic Networks

The Internet of Things technology has brought about the arrival of the era of big data.Various data sources such as sensors and the Internet generate a large amount of data at any time.How to identify the events of interest from these data becomes very important.Therefore,complex event processing has been obtained a lot of attention.In practical applications,due to noise,occlusion,and poor transmission technology,the collected event streams are often uncertain,and existing complex event processing techniques are difficult to cope with.In addition,due to insufficient understanding of the field,the complex event rules summarized from it will be inaccurate,and in the process of processing,wrong results will occur.At present,there are still many problems to be solved in the uncertain complex event processing technology.How to identify complex events in the case of inaccurate rule definition and incomplete original event streams,how to calculate the probability of an complex event occurring in the case of the original event streams with probabilistic properties.Aiming at the problems in complex event processing technology,this paper proposes Uncertainty Complex Event Processing based on Markov Logic Networks(MLN-UCEP).Based on this method,the original event streams with probabilistic properties is considered,a method for dealing with uncertain complex events based on soft evidence update(UCEP-SE)is proposed.The main work of this paper includes:(1)Studying and analyzing the uncertainty caused by dirty reading,missed reading,multi-read and the interference in data transmission in the real world,and it is classified according to the expression of uncertainty.Aiming at the incomplete flow of the original event and the inaccurate definition of the rules,based on the event calculus,a method for uncertain complex event processing based on Markov logic networks is proposed,and the predicates and axioms in event calculus is rewritten by introducing technologies such as Closed World Assumption to reduce the number of closed predicates.Experimental results show that this method is more powerful than traditional complex event processing methods.(2)Some original events come from other uncertain complex event processing systems and are accompanied by probabilistic properties.For this reason,based on MLN-UCEP,this paper proposes a method for processing complex uncertain events based on soft evidence updating.During the procedure of the Markov logic networks,the size of the ground Markov networks is reduced by calculating the irrelevant information independent of any query,and the probabilistic properties of the original event are integrated into the inference process without generating a large amount of computational overhead.