以事件为核心的面向对象时空数据模型

GIS中时态问题的解决不仅要在数学上寻求方法的支持，更应该在现实世界的时空现象中去探求其内在的变化规律。事件是驱动时空数据发生变化的根本原因，但每一次事件对数据对象的影响总是局限于一定的范围，这内在地表现为对象的属性之间的时空内聚性并决定了时态数据管理的对象粒度。时空数据模型的设计必须参照这一规律才能在冗余与效率之间取得平衡。在简要分析现有的几种典型时空数据模型的基础上，根据上述观点设计了一种以事件为核心的面向对象时空数据模型。房产信息系统的设计与实现证明，该模型对于时空数据管理效果较好，为进一步抽取变化的模式和预测决策奠定了基础。

An Event-centric Object Oriented Spatio-temporal Data Model

Rather than a mathematic approach, the intrinsic changing law of spatio-temporal phenomena of the real world should be concerned to attack the temporal problem in GIS. The event is a fundamental factor that results in data's change. But each event only affects the data within a certain range, which is intrinsically corresponds with the cohesion between object properties. And the cohesion determines the object granularity to be marked with time. To find a balance between data redundancy and system complexity, the law must be complied with in designing a spatio-temporal data model. According to this point of view, the paper analyzed several existing typical spatio-temporal data models and discussed a few correlative questions briefly. And then proposed an event-centric object-oriented spatio-temporal data model. In an application example of real estate information system, the model is proved to be good for establishing a foundation for further study of model-abstract, prediction and decision-making.