The Prediction Of Geosensor Data Based On The Coupled Spatio-temporal K-nearest Neighbors And Vector Autoregressive

With the development of the sensor network,more and more geosensor data are collected,and the prediction of geosensor data has become a hot topic among scholars.The prediction of geosensor data is widely used in economy,engineering,natural science and social sciences.At present,scholars have put forward a lot of prediction methods for geosensor data,such as STARIMA,cARIMA,cVAR,etc.These methods process the spatial correlation of different sites and the time correlation of the same site in the data in different ways.But the methods mentioned above which process data by clustering just apply the cluster result to all the sequence in a same cluster through a same way,ignored the uniqueness of each geosensor data sequence,the accuracy of geosensor data prediction remains to be improved.In this paper,a nVAR model which computes the relevance of the space-time information effectively and considers the uniqueness of each sensing sequence at the same time is proposed to predict the geosensor data.The proposed quantifies the time information and spatial information of the data by calculating the space-time distance,and then searches for the K nearest neighbor based on the space-time distance.Finally,the nearest neighbor sequences are applied to the vector autoregressive model.With the method of search for space-time nearest neighbors,nVAR model compute the relevance of the time dimension and space dimension effectively.At the same time,nVAR model uses the space-time nearest neighbor sequences which are highly correlated to predict the sensing sequence,and gives full consideration to the uniqueness of the various geographic sequence.In addition,the methods mentioned above only pay attention to the correlation between any two time sequences when dealing with time correlation,overlooked the overall situation.Based on the nVAR model,a cnVAR model was proposed.We introduced the coupling time distance in cnVAR model,processed time information by coupling,defined time distance from the global perspective.Seven data sets collected from four real geosensor networks were used to test the cVAR?nVAR and cnVAR model in the end.The experimental results show that the nVAR and cnVAR model can improve the prediction accuracy of geosensor data effectively.