Event Structuring as a General Approach to Building Knowledge in Time-Based Collections

Abstract

Many kinds of data collections are time-based or can be collected in a temporal manner. There has been a desire in the geography and geospatial communities to put temporal behavior on the same footing as spatial structure and to develop a comprehensive geo-temporal information system. In many cases, temporal information refers to sequences of happenings in the world. To efficiently represent such temporal information, we present event structuring as a general approach to build knowledge in time-based collections. In this case, an event is defined as a meaningful occurrence that has substantial impact on subsequent developments. A properly organized event sequence forms a narrative, or story. Such stories are powerful mechanisms for human understanding; not only are they in a form that make them easier to recall, but they also lead to mental models that can be intuited, examined, and joined together into larger models. In this paper, the proposed event structuring methods are not limited to geospatial data, but apply to any type of time-based collection such as text corpora. We first provide the definition of event structuring, and then describe detailed examples of event structures built upon different kinds of data. Last, we raise the need for an event descriptive language in order to generate, organize, and compare event structures.