Spatio-temporal data modeling of rich semantic for composite events
The spatio-temporal feature is a fundamental characteristic of an event.A composite event is the intricate amalgamation of multiple events that have multi-scale spatio-temporal features and multi-level semantic attributes.To effectively model and analyze composite events,it is essential to leverage a spatiotemporal data model that captures the basic features of events and to realize the knowledge mining as well as the simulation and reasoning of the spatio-temporal evolution of events.Existing models often struggle to portray the evolutionary trajectories and causal relationships between entities across different scales of composite events,and lack the ability to capture the dynamically semantic changes of composite events,hindering their ability to meet the requirements for accurate modeling.The primary contributions of this paper are as follows:(1)By analyzing the concept of events within composite events,we propose a rich semantic event spatio-temporal data model that broadens the dimensions of events covered by existing models.We describe the diverse states and behaviors associated with the evolution of multiple entities across various temporal and spatial granularities based on the multidimensional characteristics of events in different domains.(2)We aggregate event hierarchies with their characteristics to establish a multi-level logical framework of events,progressing from complexity to simplicity,from the event itself to its constituent elements.(3)In our model,we depict the associations between events and events,as well as entities and entities within complex events,and analyze the potential causal relationship of the event through the evolution of spatio-temporal relationships and attribute relationships.(4)Finally,we elaborate on the modeling and analysis methods of the rich semantic event spatio-temporal data model by instantiating a historical war in human society.We illustrate the model's application using a GIS platform,demonstrating the knowledge inference process of evolving entity states and changing relationships between entities triggered by events,which lead to the occurrence of new events.Through this process,we validate the feasibility and practicality of our model.
composite eventrich semanticevent spatio-temporal data modelmultiple time and space scalesspatio-temporal evolution causal relationships
万方数据
维普
