资 源 简 介
We proposed a framework called spatiotemporal zoning as an attempt to overcome the limitations of geo-temporal encoding faced in the existing health surveillance systems by classifying news content into predefined classes based on its spatial and temporal characteristic, and recognizing the spatial and temporal attributes of each event.
Specifically speaking, spatiotemporal zoning is the task that aims to partition text into segments that contain events, which occurred in the same location at the same homogeneous time frame. Here the homogeneous time frame means events in the text segment are overlap in time, continually or sequentially occurred, or strongly relate in time.
