A novel frequent patterns mining method of unusual climate events in data of East Asian monsoon zone

Unusual climate events which may cause disasters have great influence on both the natural environment and the human society. Finding association patterns among these events has great significance. Traditional data mining methods have several problems while applied to climate science data directly so we propose a novel method that mining frequent patterns among unusual events in climatic data, including spatial clustering algorithm based on tight clique, extracting unusual climate events algorithm and extended generalized sequential pattern (EGSP) algorithm. In order to verify our method, we do experiments on real climatic data (Climatic data of East Asian monsoon zone) and find lots of well-known and previously unknown patterns. It needs the climatic expert to judge whether the new patterns are significative. Overall, the experiment told us it was an effective and viable method for the climatic research.