海陆气候事件关联规则挖掘方法
收稿日期: 2013-06-08
修回日期: 2013-07-06
网络出版日期: 2014-03-10
基金资助
教育部新世纪优秀人才资助计划（NECT-10-0831）；高等学校博士学科点专项科研基金项目（20110162110056）；江苏省资源环境重点实验室开放基金项目（JS201101）；现代工程测量国家测绘地理信息局重点实验室开放基金项目（TJES1102）。
Discovering Sequential Association Rules between Single Ocean Climate Index and Land Abnormal Climate Events
Received date: 2013-06-08
Revised date: 2013-07-06
Online published: 2014-03-10
近年来，异常气候事件的频发对人类的生活环境和经济发展带来严重负影响。气象学家研究表明：海洋气候异常对陆地气候异常事件的发生具有重要的诱发作用，因此，对海陆气候间的内在关联机制进行深入挖掘具有重要研究价值。本文提出了一种关联规则挖掘方法，以探索单一海洋气候指数与陆地异常气候事件间存在的关联。首先，针对陆地气候要素，采用顾及空间邻近关系的层次聚类方法进行有效气候分区，通过对各层分区结果进行相关统计分析得到有效的各区域气候序列；然后，进行顾及多重约束进行时序关联规则挖掘，以探索海陆气候要素间的关联机制；最后，通过实际算例分析得到的各气候指数与我国陆地区域异常降水事件间的关联机制结果，与实际情况高度吻合。
石岩, 邓敏, 刘启亮, 杨文涛 . 海陆气候事件关联规则挖掘方法[J]. 地球信息科学学报, 2014 , 16(2) : 182 -190 . DOI: 10.3724/SP.J.1047.2014.00182
With the frequent occurrence of abnormal climatic events in recent years, social economic and people's life are impacted more and more seriously. Meteorologists have found that ocean climate has an effect on land climate, such that the EI NINO can lead abnormal precipitation events on some land regions. Therefore, it is very critical to study the associations between ocean and land climate factors. At present, some researchers have done a series of work about this aspect and several representative methods have been proposed. The eigenvalue statistics and traditional sequential association rules mining are two main methods. However, the former is sensitive to noise and not suitable for huge amounts of data, while the latter dose not fully consider the correlation and multi-scale properties hidden in the climate time series data. In view of this, a method based on multi-constraints is proposed to discover sequential association rules between individual ocean and land climate factors in this paper. First, we took both time correlation and spatial correlation into account and a hierarchical clustering method with the consideration of spatial proximity is employed to find climate zones for the land climate factor. In this way, we not only preserve the effective information in the data, but also make the raw data simpler by removing time correlation and spatial correlation. Second, the land and ocean climate sequences are discretized based on domain knowledge and a series of events are also extracted. These are further used to construct the transactions mining table. Finally, a new method, which utilizes multiple constraints, is developed to mine sequential association rules. We only focus on the associations between ocean climate indices and abnormal land climate events, such as flood and drought. As a matter of fact, we need the frequent rules which can describe a law to a certain extent. A practical example is used to explore the relationships between each climate index and unusual precipitation events in China, and the results obtained are very consistent with the actual situation. This to a large degree illustrates that the method proposed in this paper is rational. In addition, we also gain some unknown knowledge that can provide some information for meteorologists. Based on the information, meteorologists can study the internal mechanism deeply. Also, the information can guide the government to make related policy decisions. In summary, the method in this paper takes the spatial correlation, time correlation and multi-scale characteristics into account effectively, while considers multi-constraint to deal with climate problems more accurately. By experiments, it is proved that our method is correct and valid.
[1] 姜世中.气象学与气候学[M].北京:科学出版社,2010.
[2] Tan P, Steinbach M, Kumar V, et al. Finding spatio-temporal patterns in earth science data[C]. Proceedings of KDD Workshop on Temporal Data Mining, San Francisco, U.S.A, 2001.
[3] Wold S. Principal component analysis[J]. Chemometrics and Intelligent Laboratory Systems, 1987, 2(1): 37-52.
[4] Klema V, Laub A. The singular value decomposition: Its computation and some applications[J]. IEEE Transactions on Automatic Control, 1980, 25(2): 164-176.
[5] Han J W, Kamber M. Data mining: Concepts and technique[M]. San Francisco: Morgan Kaufmann, 2005.
[6] 邓敏,刘启亮,李光强,等.空间聚类分析及应用[M].北京:科学出版社,2011.
[7] Fovell R, Fovell M. Climate zones of the conterminous United States defined using cluster analysis[J]. Journal of Climate, 1993,6(11),2103-2135.
[8] Fovell R. Consensus clustering of U.S. temperature and precipitation data[J]. Journal of Climate, 1997,10(6):1405-1427.
[9] Agarwal R, Srikant R. Fast algorithms for mining association rules[C]. Proceeding of the 20th International Conference on Very Large Databases, 1994,487-499.
[10] Agarwal R, Srikant R. Mining sequential patterns[C]. Proceedings of the 11th International Conference on Data Engineering, 1995,3-14.
[11] Mannila H, Toivonen H, Verkanmo A. Discovery of frequent episodes in event sequences[J]. Data Mining and Knowledge Discovery, 1997,1(3):259-289.
[12] Das G, Lin K I, Mannila H, et al. Rule discovery from time series[C]. Proceedings of the 4th International Conference on Knowledge Discovery and Data Mining, New York, U.S.A, 1998, 16-22.
[13] Harms S K, Deogun J, Tadesse T. Discovering sequential association rules with constrains and time lags in multiple sequences[C]. Proceedings of the 2002 International Symposium on Methodologies for Intelligent Systems, Lyon, France, 2002,431-441.
[14] 王佳璆,邓敏,程涛,等.时空序列数据分析和建模[M].北京:科学出版社,2012.
[15] Cheng T, Wang J Q, Li X. A Hybrid framework for space-time modeling of environmental data[J]. Geographical Analysis, 2011,43(3):188-210.
[16] 孔令桥,秦昆,龙腾飞.利用二型模糊聚类进行全球海表温度数据挖掘[J].武汉大学学报(信息科学版),2012,37(2):215-219.
[17] Wu T, Song G, Ma X J, et al. Mining geographic episode association patterns of abnormal events in global earth science data[J]. Science in China Series E: Technological Sciences, 2008,51(1):155-164.
[18] Lin F, Jin X X, Hu C, et al. Discovery of teleconnections using data mining technologies in global climate datasets[J]. Data Science Journal, 2007,6(17):749-755.
[19] Deng M, Liu Q L, Cheng T, Shi Y. An adaptive spatial clustering algorithm based on Delaunay triangulation[J]. Computers, Environment and Urban Systems, 2011,35(4):320-332.
[20] Guo D. Regionalization with dynamically constrained agglomerative clustering and partitioning (REDCAP)[J]. International Journal of Geographical Information Science, 2008,22(7):801-823.
[21] Tadesee T, Wilhite D A, Harms S K, et al. Drought monitoring using data mining techniques: A case study for Nebraska, USA[J]. Natural Hazards, 2004,33(1):137-159.
[22] Bezdek J C, Nikhil R P. Some new indexes of cluster validity[J]. IEEE Transactions on Systems, Man, and Cybernetics-Part B, 1998,28(3):307-310.
/
| 〈 |
|
〉 |