基于主成分分析的支持向量机划分地震相-应用数学专业论文.docxVIP

Subject Subject ：Seismic Facies Classification by Support Vector Machine Based on Principal Component Analysis Specialty ：Applied Mathematics Name ：Zhu Yanwei (signature)星丛i血丝 Instructor：Liu Yeling (Signature) ABSTRACT In this thesis，seismic facies was classed by support vector machine based on principal component analysis．SVM based on principal component analysis classed seismic facies， which speeded velocity of仃mmng and forecast up，maintained the classified information of samples and not changed distribution of samples，and so enhance the seismic facies class precision．Accuracy of classification seismic facies affected directly result’s reliability of the oil gas exploration and forecast，helped to reduce exploration’s risk，and brought the great efficiency of economy and the society． In the paper，the related knowledge of support vector machine was introduced，the algorithm of realization support vecmr machine was discussed，especially sequential minimal optimization algorithm was analyzed，and the advantage and characteristic of this optimized algorithm was summarized comparing chunking algorithm and the decomposing algorithm． Regarding algorithm of SVM for multi—class classification，the advantage and the disadvantage of one—against—one and one—against—the—rest were summarized，and coding multi-class classifications algorithm based on PCA support vector machine was proposed． The step and forecast process of the algorithm welTe explained．Through the trial of seismic data,it was proved that the proposed algorithm in the paper reduced the complication of calculation，saved the time of calculation，and increased classification accuracN Using this algorithm to class seismic facies，not only the classification speed is fast but also forecast precision obviously is high．In addition，during classification of seismic facies，the method of dam partition was adopted the problem of enormous seismic data． At last，throu曲trail，regarding the