The Methods and Key Issues of Data Mining on Landslide

摘要

After obtained a large number of low-altitude remote sensing data of hazards in Wenchuan earthquake stricken area, how to quickly extract the disaster information and access the hazards law became the key issue that restricted in post-disaster reconstruction. Because of Landslide data's characteristics of highly nonlinear, fuzzy, great data volume, type's diversity; introduction of data mining and efficient intelligent information technologies becomes a necessary requirement. Data mining can quickly obtain landslide quantity, types, volume, distribution and other basic features and information; it also can predict and analyze space-time evolution characteristics of hazards chain, that fully reflects the advantage of application of data mining in the field of disaster. Based on the basic characteristics of landslide data and data mining methods can be used, The paper establishes a landslide information mining system containing landslide database, data mining module and landslide information module. Then typical analysis on the landslide data mining system is made by a example of landslide risk assessment. At last, the key problems in data mining on landslide information are proposed for the present research. The conclusions can provide a reference for data mining technology promoting in the field of disaster.