基于量子蒙特卡罗的地震多属性聚类方法
|
摘要
多属性聚类分析是从众多地震属性中提取地下地质特征的重要途径。为了更好地提高属性聚类分析的有效性,本文将寻优能力更强的量子蒙特卡罗方法引入到聚类分析中,它能自动根据数据结构的特点,动态调整数据固有的类别数;通过在聚类分析过程中引入相关性分析,估算各个属性的权重大小,提出根据权重大小进行加权的方法,突出对地质特征敏感的地震属性的作用;并提出变尺度的方法,挖掘属性之间共有的、宏观的特征之外的细节部分,减少属性交叉信息的影响。实例应用结果表明,本文提出的方法能很好地挖掘数据的内在特征,提高储层预测的准确性。
Multi-attributes clustering is an important way for drawing underground geologic features from a large number of seismic attributes.For improving the effectiveness of the attributes clustering,this paper introduces quantum Monte Carlo method with stronger optimization ability into the clustering to adjust the inherent category number of data dynamically according to the characteristics of the data structure.By increasing the correlation analysis in the process of clustering for estimating each attribute weight,seismic attributes which are sensitive to geological features can be highlighted.The proposed variable scale method can show details together with macro common attribute characteristics to reduce the influence of attribute cross information.Examples show that the method proposed in this paper could be very good at mining the inherent characteristics of data to improve reservoir prediction accuracy.
Multi-attributes clustering is an important way for drawing underground geologic features from a large number of seismic attributes.For improving the effectiveness of the attributes clustering,this paper introduces quantum Monte Carlo method with stronger optimization ability into the clustering to adjust the inherent category number of data dynamically according to the characteristics of the data structure.By increasing the correlation analysis in the process of clustering for estimating each attribute weight,seismic attributes which are sensitive to geological features can be highlighted.The proposed variable scale method can show details together with macro common attribute characteristics to reduce the influence of attribute cross information.Examples show that the method proposed in this paper could be very good at mining the inherent characteristics of data to improve reservoir prediction accuracy.
引文
[1]Chen Q,Sidney S.Seismic attribute technology forreservoir forecasting and monitoring.The LeadingEdge,1997,16(5):445～456
[2]印兴耀,周静毅.地震属性优化方法综述.石油地球物理勘探,2005,40(4):482～489Yin Xingyao,Zhou Jingyi.Summary of optimummethods of seismic attributes.OGP,2005,40(4):482～489
[3]曹辉.关于地震属性应用的几点认识.勘探地球物理进展,2002,25(5):18～22Cao Hui.Discussion on the application of seismic at-tributes.Progress in Exploration Geophysics,2002,25(5):18～22
[4]Bezdek J C.Pattern Recognition with Fuzzy Objec-tive Function Algorithms.Plenum Press,New York,1981,95～107
[5]Selim S Z,Ismail M A.K-means-type algorithm:Ageneralized convergence theorem and characterizationof local optimality.IEEE Trans on Pattern Analysisand Machine Intelligence,1984,6(1):81～87
[6]Brown D,Huntley C.A practical application of simu-lated annealing to clustering.Pattern Recognition,1992,25(4):401～412
[7]Mitchell T.Machine Learning.McGraw-Hill,NewYork,1997
[8]Bradley P,Fayyad U.Refining initial points for k-means clustering.Proceedings of the 15th ICML,Madison,1998,91～99
[9]Pelleg D,Moore A.X-means:Extending K-meanswith efficient estimation of the number of the clus-ters.Proceedings of the 17th ICML,2000
[10]Strehl A,Ghosh J.A scalable approach to balanced,high dimensional clustering of market baskets.Pro-ceedings of the 17th International Conference onHigh Performance Computing,Bangalore:SpringerLNCS,2000,525～536
[11]Zhang B.Generalized K-harmonic means:Dynamicweighting of data in unsupervised learning.Proceed-ings of the 1st SIAM ICDM,Chicago,2001
[12]Kohonen T.Self-organizing Maps.Springer Seriesin Information Sciences,Springer,2001
[13]Dhillon I,Guan Y,Kogan J.Refining clusters in highdimensional data.The 2nd SIAM ICDM Workshopon Clustering High Dimensional Data,Arlington,2002
[14]Sarafis I,Zalzala A M S,Trinder P W.A genetic rule-based data clustering toolkit.Congress on Evolution-ary Computation,Honolulu,2002
[15]Banerjee A,Ghosh J.On scaling up balanced cluste-ring algorithms.Proceedings of the 2nd SIAM IC-DM,Arlington,2002
[16]Berkhin P,Becher J.Learning simple relations:Theo-ry and applications.Proceedings of the 2nd SIAMICDM,Arlington,2002,333～349
[17]Cao Y Q,Wu J H.Dynamics of projective adaptiveresonance theory model:The foundation of PART al-gorithm.IEEE Transactions on Neural Network,2004,15(2):245～260
[18]Ester M,Kriegel H P,Sander J et al.A density-basedalgorithm for discovering clusters in large spatial da-tabases with noise.Proceedings of the 2nd ACMSIGKDD,Portland,1996,226～231
[19]Wang W,Yang J,Muntz R.Sting:A statistical infor-mation grid approach to spatial data mining.Proceed-ings of the 23rd Conference on VLDB,Athens,1997,186～195
[20]Guha S,Rastogi R,Shim K.Cure:An efficient cluste-ring algorithm for large databases.Proceedings ofthe ACM SIGMOD Conference,Seattle,1998,73～84
[21]Guha S,Rastogi R,Shim K.Rock:A robust cluste-ring algorithm for categorical attributes.Proceedingsof the 15th ICDE,Sydney,1999,512～521
[22]Hinneburg A,Keim D.An efficient approach to clus-tering large multimedia databases with noise.Pro-ceedings of the 4th ACM SIGKDD,New York,1998,58～65
[23]Agrawal R,Gehrke J,Gunopulos D et al.Automaticsubspace clustering of high dimensional data for datamining applications.Proceedings of the ACM SIG-MOD Conference,Seattle,1998,94～105
[24]程乾生.属性均值聚类.系统工程理论与实践,1998,18(9):124～126Cheng Qiansheng.Attribute means clustering.Sys-tems Engineering—Theory&Practice,1998,18(9):124～126
[25]Karypis G,Han E H,Kumar V.Chameleon:A hie-rarchical clustering algorithm using dynamic model-ing.IEEE Computer,1999,32(8):68～75
[26]Ertoz L,Steinbach M,Kumar V.Finding clusters ofdifferent sizes,shapes and densities in noisy,high di-mensional data.Proceedings of the 3rd SIAM Inter-national Conference on Data Mining,2003,112:47～59
[27]武兆慧,张桂娟,刘希玉.基于模拟退火遗传算法的聚类分析.计算机应用研究,2005,22(12):24～26Wu Zhaohui,Zhang Guijuan,Liu Xiyu.Clusteringbased on simulated annealing genetic algorithm.Ap-plication Research of Computers,2005,22(12):24～26
[28]董金新,亓民勇.一种新的基于粒子群和模拟退火的聚类算法.计算机工程与应用,2009,45(35):139～142Dong Jinxin,Qi Minyong.New clustering algorithmbased on particle swarm optimization and simulatedannealing.Computer Engineering and Applications,2009,45(35):139～142
[29]陈波,胡少华,毕建军.地震属性模式聚类预测储层物性参数.石油地球物理勘探,2005,40(2):204～208Chen Bo,Hu Shaohua,Bi Jianjun.The petrophysicalparameters prediction by seismic attribute clustering.OGP,2005,40(2):204～208
[30]杨培杰,印兴耀,张广智.模糊C均值地震属性聚类分析.石油地球物理勘探,2007,42(3):322～325Yang Peijie,Yin Xingyao,Zhang Guangzhi.FuzzyC-mean seismic attribute clustering analysis.OGP,2007,42(3):322～325
[31]Donsker M D,Kac M.A sampling method for deter-mining the lowest eigenvalue and the principal eigen-function of Schrdinger equation.J Res Natl BurStd,1950,44:551～557
[32]魏超,李小凡,张美根.基于量子蒙特卡罗的地球物理反演方法.地球物理学报,2008,51(5):1494～1502Wei Chao,Li Xiaofan,Zhang Meigen.The geophysi-cal inverse method based on quantum Monte Carlo.Chinese Journal of Geophysics,2008,51(5):1494～1502
[33]马文淦.计算物理学.北京:科学出版社,2005
[34]Pang T.An Introduction to Computational Physics.Cambridge University Press,2006
[2]印兴耀,周静毅.地震属性优化方法综述.石油地球物理勘探,2005,40(4):482～489Yin Xingyao,Zhou Jingyi.Summary of optimummethods of seismic attributes.OGP,2005,40(4):482～489
[3]曹辉.关于地震属性应用的几点认识.勘探地球物理进展,2002,25(5):18～22Cao Hui.Discussion on the application of seismic at-tributes.Progress in Exploration Geophysics,2002,25(5):18～22
[4]Bezdek J C.Pattern Recognition with Fuzzy Objec-tive Function Algorithms.Plenum Press,New York,1981,95～107
[5]Selim S Z,Ismail M A.K-means-type algorithm:Ageneralized convergence theorem and characterizationof local optimality.IEEE Trans on Pattern Analysisand Machine Intelligence,1984,6(1):81～87
[6]Brown D,Huntley C.A practical application of simu-lated annealing to clustering.Pattern Recognition,1992,25(4):401～412
[7]Mitchell T.Machine Learning.McGraw-Hill,NewYork,1997
[8]Bradley P,Fayyad U.Refining initial points for k-means clustering.Proceedings of the 15th ICML,Madison,1998,91～99
[9]Pelleg D,Moore A.X-means:Extending K-meanswith efficient estimation of the number of the clus-ters.Proceedings of the 17th ICML,2000
[10]Strehl A,Ghosh J.A scalable approach to balanced,high dimensional clustering of market baskets.Pro-ceedings of the 17th International Conference onHigh Performance Computing,Bangalore:SpringerLNCS,2000,525～536
[11]Zhang B.Generalized K-harmonic means:Dynamicweighting of data in unsupervised learning.Proceed-ings of the 1st SIAM ICDM,Chicago,2001
[12]Kohonen T.Self-organizing Maps.Springer Seriesin Information Sciences,Springer,2001
[13]Dhillon I,Guan Y,Kogan J.Refining clusters in highdimensional data.The 2nd SIAM ICDM Workshopon Clustering High Dimensional Data,Arlington,2002
[14]Sarafis I,Zalzala A M S,Trinder P W.A genetic rule-based data clustering toolkit.Congress on Evolution-ary Computation,Honolulu,2002
[15]Banerjee A,Ghosh J.On scaling up balanced cluste-ring algorithms.Proceedings of the 2nd SIAM IC-DM,Arlington,2002
[16]Berkhin P,Becher J.Learning simple relations:Theo-ry and applications.Proceedings of the 2nd SIAMICDM,Arlington,2002,333～349
[17]Cao Y Q,Wu J H.Dynamics of projective adaptiveresonance theory model:The foundation of PART al-gorithm.IEEE Transactions on Neural Network,2004,15(2):245～260
[18]Ester M,Kriegel H P,Sander J et al.A density-basedalgorithm for discovering clusters in large spatial da-tabases with noise.Proceedings of the 2nd ACMSIGKDD,Portland,1996,226～231
[19]Wang W,Yang J,Muntz R.Sting:A statistical infor-mation grid approach to spatial data mining.Proceed-ings of the 23rd Conference on VLDB,Athens,1997,186～195
[20]Guha S,Rastogi R,Shim K.Cure:An efficient cluste-ring algorithm for large databases.Proceedings ofthe ACM SIGMOD Conference,Seattle,1998,73～84
[21]Guha S,Rastogi R,Shim K.Rock:A robust cluste-ring algorithm for categorical attributes.Proceedingsof the 15th ICDE,Sydney,1999,512～521
[22]Hinneburg A,Keim D.An efficient approach to clus-tering large multimedia databases with noise.Pro-ceedings of the 4th ACM SIGKDD,New York,1998,58～65
[23]Agrawal R,Gehrke J,Gunopulos D et al.Automaticsubspace clustering of high dimensional data for datamining applications.Proceedings of the ACM SIG-MOD Conference,Seattle,1998,94～105
[24]程乾生.属性均值聚类.系统工程理论与实践,1998,18(9):124～126Cheng Qiansheng.Attribute means clustering.Sys-tems Engineering—Theory&Practice,1998,18(9):124～126
[25]Karypis G,Han E H,Kumar V.Chameleon:A hie-rarchical clustering algorithm using dynamic model-ing.IEEE Computer,1999,32(8):68～75
[26]Ertoz L,Steinbach M,Kumar V.Finding clusters ofdifferent sizes,shapes and densities in noisy,high di-mensional data.Proceedings of the 3rd SIAM Inter-national Conference on Data Mining,2003,112:47～59
[27]武兆慧,张桂娟,刘希玉.基于模拟退火遗传算法的聚类分析.计算机应用研究,2005,22(12):24～26Wu Zhaohui,Zhang Guijuan,Liu Xiyu.Clusteringbased on simulated annealing genetic algorithm.Ap-plication Research of Computers,2005,22(12):24～26
[28]董金新,亓民勇.一种新的基于粒子群和模拟退火的聚类算法.计算机工程与应用,2009,45(35):139～142Dong Jinxin,Qi Minyong.New clustering algorithmbased on particle swarm optimization and simulatedannealing.Computer Engineering and Applications,2009,45(35):139～142
[29]陈波,胡少华,毕建军.地震属性模式聚类预测储层物性参数.石油地球物理勘探,2005,40(2):204～208Chen Bo,Hu Shaohua,Bi Jianjun.The petrophysicalparameters prediction by seismic attribute clustering.OGP,2005,40(2):204～208
[30]杨培杰,印兴耀,张广智.模糊C均值地震属性聚类分析.石油地球物理勘探,2007,42(3):322～325Yang Peijie,Yin Xingyao,Zhang Guangzhi.FuzzyC-mean seismic attribute clustering analysis.OGP,2007,42(3):322～325
[31]Donsker M D,Kac M.A sampling method for deter-mining the lowest eigenvalue and the principal eigen-function of Schrdinger equation.J Res Natl BurStd,1950,44:551～557
[32]魏超,李小凡,张美根.基于量子蒙特卡罗的地球物理反演方法.地球物理学报,2008,51(5):1494～1502Wei Chao,Li Xiaofan,Zhang Meigen.The geophysi-cal inverse method based on quantum Monte Carlo.Chinese Journal of Geophysics,2008,51(5):1494～1502
[33]马文淦.计算物理学.北京:科学出版社,2005
[34]Pang T.An Introduction to Computational Physics.Cambridge University Press,2006
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心