Mining high-temperature event space-time regions in geo-referenced temperature series data

Mining space-time regions of events is an important task in data mining. It has wide applications in various disciplines, such as epidemiology, meteorology. The existing space-time regions events mining algorithms usually based on clustering analysis, which is difficult to detect irregularly shaped events when they evolve by time. Meanwhile parameter-setting is also a difficult problem for most existing methods. In this paper, we propose a novel automatic event mining algorithm-Gtem. Combined with Minimum Length Description (MDL) principle, Gtem can optimize parameter-setting; detect event regions of different evolutions according to the spatial-temporal correlations of objects and find outliers as well. We conduct experiments on daily-weather datasets of Hunan province from 2004-2008 and the experimental results show that the proposed Gtem could find high-temperature space-time regions efficiently.