Research on Knowledge Discovery on Documents in Spatiotemporal Document Streams

点击翻译按钮获取中文摘要

This study has developed a novel method for extracting spatiotemporal social events and hot topics in a spatiotemporal document stream. In this study, a mathematical model for spatiotemporal document stream is defined. Spatiotemporal social events and hot topics can be extracted by using the location-based burst detection algorithm. To extract bursts for topics, the clustering-based burst detection algorithm is proposed. Moreover, the parallelization method for burst detection algorithm, which is developed in this study, archives the speed-up of extracting bursts on the large-scale spatiotemporal document streams.