论文部分内容阅读
数据挖掘是从大量的、不完全的、有噪声的、模糊的数据中提取隐含的、但又是潜在有用的信息和知识的过程,由于数据挖掘具有出色的非线性建模能力和自组织学习能力,因此可以在复杂储层的测井解释中发挥作用。本文用数据挖掘方法识别复杂储层的岩性。将岩性识别作为一种分类任务建立数据挖掘流程,包括特征提取、特征选择和建立模型等步骤。本文用独立成分分析法从测井曲线中提取信息;然后使用分支定界算法寻找最佳的特征子集,并消除冗余信息;最后采用C5.0决策树算法建立分类模型的测井曲线。模型和实际测井数据吻合较好,表明在复杂油藏的研究中数据挖掘方法是有效的。
Data mining is the process of extracting implicit but potentially useful information and knowledge from a large amount of incomplete, noisy and obscure data. Because data mining has excellent nonlinear modeling ability and self-organization Learning ability, therefore, can play a role in logging interpretation of complex reservoirs. In this paper, data mining methods to identify the lithology of complex reservoirs. The lithology identification as a classification task to establish data mining process, including feature extraction, feature selection and modeling steps. In this paper, we use independent component analysis to extract information from well logs. Then, we use branch and bound algorithm to find the best subset of features and eliminate redundant information. Finally, the C5.0 decision tree algorithm is used to establish the logging curves of classification models. The agreement between the model and the actual well logging data is good, which shows that the data mining method is effective in the study of complex reservoirs.