【摘要】 地震属性数据中包含着一些潜在的有用信息,但从地震资料中提取的地震属性多达数百种,这对实际研究造成很大困难。根据地震属性之间必然存在着相关性这一前提,对地震属性进行降维分析是解决这一困难的有效措施。将数据挖掘技术中的K-L变换和奇异值分解2种线性降维方法,分别应用到地震属性降维中,再将降维后的地震属性应用到研究区的实际地震数据中进行测试,结果发现降维后的地震属性均集中在特征值最大的那个新属性上。并依次减小,认为用降维后的地震属性预测储层比用单一地震属性更符合地质规律。更多还原

【Abstract】 Several hundreds of seismic attributes from seismic data make it difficult to apply in practice although the seismic attributes contain much potential useful information.Based on the interdependeney of the seismic attributes,the reduction of dimensions of seismic attributes is a valid method to resolve this problem.Here we use the K-L transformation and SVD methods in the reduction of dimensions of seismic attributes.The test of practical data indicates that the main features in the seismic attributes are distributed in the first main component after dimension reduction,and decrease successively.It is better to predict reservoirs by using the seismic attributes reduced dimensions than single seismic attribute.更多还原