明代雹灾的关联信息挖掘研究

The Mining Research on the Association Information of Hail Diasters in the Ming Dynasty

对收集的明代雹灾史料进行整理和分类,构建了我国明代的雹灾数据库,将雹灾的定性描述转化成定量描述。根据冰雹的大小和造成的危害进行等级划分,对发生雹灾时伴随的其它阵性天气现象进行统计分析。利用关联挖掘模型continuous association rule mining algorithm(CARMA)算法挖掘不同等级的雹灾与风、雨、雷电等阵性天气的关联性及其有关规律。结果表明:明代的雹灾与雨的关联性最大,其次是风和雷电;它们易带来大雨雹或大风雨雹等灾害。同时不同等级的雹灾发生时伴随着其他的阵性天气的置信度也存在着差异;因此当雹灾出现的时候,应同样考虑到其他如暴雨、狂风、雷电等强对流天气灾害造成的危害。获得的雹灾关联信息,可用于指导雹灾的预报与支持决策。

A great amount of historical data of hailstorm in Ming dynasty were collected,sorted and classified.First of all,the hailstorm situations of Ming dynasty were established database,and the qualitative descriptions of hailstorm were converted into quantitative descriptions.Hailstorm was divided into different hierarchies depending on the size and damage of hail.A statistical analysis of weather phenomena associated with the hail disaster was made.Associated mining model ——CARMA algorithm was applied on mining the rules between hail and other gustiness weathers,like wind,rain,lighting,at different levels.The results in Ming Dynasty show that：14 association rules are discovered between three different levels of hail disasters with other gustiness weathers,the greatest relevance is hail and rain,the second is the wind,they are easy to bring heavy rain or storm hail disasters.The relevance of different levels of hail disasters with other gustiness weathers also exist differences.The meaningful hailstorm correlation patterns is obtained to guide the hailstorm forecast and support decision-making.