利用独立成分分析和朴素贝叶斯算法识别含硬石膏储层的岩性  

Lithology identification of reservoir containing anhydrite based on independent component analysis and Naive Bayes algorithm

作　　者：石宁[1,2] 李洪奇[1,2] 罗伟平[1,2] 

机构地区：[1]中国石油大学(北京)地球物理与信息工程学院,北京102249 [2]油气数据挖掘北京市重点实验室,北京102249

出　　处：《西安石油大学学报（自然科学版）》2013年第5期39-42,4,共4页Journal of Xi’an Shiyou University（Natural Science Edition）

基　　金：国家科技重大专项(编号:2011ZX05023-005-006)

摘　　要：提出了利用独立成分分析结合朴素贝叶斯算法识别岩性的方法,并以某含硬石膏储层为例进行了岩性识别.结果表明,通过与独立成分分析的结合,贝叶斯算法能提高识别岩性的准确度.与单独使用贝叶斯算法相比,二者结合可使识别含硬石膏粉砂岩的正确率提高29个百分点.同时指出,在测井解释的过程中不应忽略算法的应用条件,通过合理的预处理方法,可以使解释过程更加严谨,从而提高测井解释水平.A lithology identification method for reservoirs containing anhydrite is proposed,which combines independent component analysis with Naive Bayes algorithm. And it is applied to a case,the result shows that by combining with independent component analysis,the lithology identification accuracy of Naive Bayes algorithm is greatly improved. Comparing with only using Naive Bayes algorithm,by combining with independent component analysis,the lithology identification accuracy of Naive Bayes algorithm to the siltstone containing anhydrite can be improved by about 29 percentage points. It is pointed out that in the logging interpretation,the application conditions of the used algorithm should not be ignored,the logging interpretation will be more rigorous and logging interpretation result will be more accurate by rational pretreatment.

关 键 词：岩性识别 含石膏储层 朴素贝叶斯算法 独立成分分析 

分 类 号：P618.13[天文地球—矿床学]