Spatial-temporal Data Mining Algorithm And Application Research

Spatial-temporal data mining as one of data mining research front and one ofthe hotspots, the core and foundation of its research is spatial-temporal data model,therefore the establishment and research of spatial-temporal data model are ofgreat significance. The existing spatial-temporal data model in the treatment ofspatial-temporal data each have advantages and disadvantages, but all promote themodel’s research and which’s perfection. The innovation of this paper lies in usingthe improved decision tree method to deal with spatial data, combining with thetemporal correlation theory deal with the time-varying data, and get aspatial-temporal data mining algorithm model based on Space decision tree andtime interrelated.At first this paper briefly introduced the formation and development of thespatial-temporal data, then the domestic and international research dynamic, thepurposes of the research and the main work in this paper.The second chapter, illustrated some related concepts about spatial-temporaldata model, detailed introduction and analysis of the space-time cube model,continuous snapshot model, ground state correction model, time and spacecomposite model for spatial-temporal data model with its advantages anddisadvantages. Points out that the object-oriented spatial-temporal data model in amore natural way to complicated spatial-temporal information modeling, and itcomes to be one effective way to support complex object modeling. The third chapter, discussed the basic ideas of the spatial decision treealgorithm, learning strategy and its suspensive condition of the algorithm recursiveoperation; Put forward the modified space decision tree algorithm and gave aninformation gain formula with space factor.The fourth chapter: gave an outline of correlativity, described the basiccontent of the correlation analysis. correlation coefficient calculation andcorrelation judgment, the establishment of the linear regression equation andcoefficient solution.Chapter Five: based on the national Olympic sports center’s smallsupermarket passenger flow statistics table, the Olympic venues in smallsupermarket site profile and according to the background trend line theory whichgave the2007-2009entry passenger number list, based on improved spatialdecision tree and time correlation theory give the spatial-temporal data model, andgives the expression of the model.Finally, the work in this paper was summarized, expounded the furtherresearch problems about the spatial-temporal data model and gave a futuredevelopment direction.