位置位图:海量时空数据的预处理模型与实现

Location Bitmaps:Model and Implementation of Massive Spatio-temporal Data Preprocessing

时空轨迹大数据蕴含着丰富的信息,值得深入挖掘探索。对海量时空轨迹数据的有效挖掘、利用依赖于完善的数据预处理模型。为解决多源融合、误差容错和性能优化等挑战,本文提出了位置位图模型。位置位图模型基于抽象的离散时空观建立概念构件体系,并在概念构件体系中定义运算算子体系,从而形成完整的时空预处理能力。本文还提出了位置位图模型在实际数据库系统中的具体的编码实现方法,该方法包括位区编码、时段选择、位串编码、运算实现和位图数据维护等环节。位置位图模型被用于改造真实的“经纬”时空大数据分析系统,改造效果表明位置位图模型不仅能够改善业务开发难度,还可以优化计算性能,降低存储开销。

discrete spatio-temporal view to establish a conceptual component system and defines an operation operator system in the conceptual component system,which can be expanded into a complete spatio-temporal preprocessing capability.This paper also proposes a specific encoding implementation method of the location bitmap model in practical database systems,which includes bit region coding,time period selection,bit string coding,operation implementation,and bitmap data maintenance.The location bitmap model is used to transform a real“Jingwei”spatio-temporal big data analysis system,and the transformation effect proves that the bitmap model can not only improve business development difficulty but also optimize computing performance and reduce storage costs.