Study of droughts and floods predicting system based on Spatial-temporal Data Mining

Flood/Drought disasters are the most frequent natural disasters in the world. In order to improve the skill of spatial-temporal forecasts for drought and flood and identify the most important hazard factors, a data mining of precipitation in spatial and temporal was executed in this paper. According to standardized precipitation index (SPI) based on the observed monthly precipitation data of 160 meteorological stations in P. R. China from 1951 to 2010, the data mining applied Principal Component Analysis Methods to identify hazard factor, and used Extended Empirical Orthogonal Function method to digest the information of the spatial-temporal distribution characteristics for spatiotemporal forecasting. The results present that these method have an optimum combination of spatial and temporal data that can be used to extract more data information and increase predicting accuracy for drought and flood in spatial-temporal scale.