论文部分内容阅读
讨论了储层识别中引入神经网络方法时的几个关键问题,如特征值优化、样本分析与选取及模糊聚类和大规模样本集合处理等,使得所建系统更符合实际环境,已在实践中得以证明.
Several key problems in the process of neural network identification in reservoir identification are discussed, such as eigenvalue optimization, sample analysis and selection, fuzzy clustering and large-scale sample set processing, which make the system more suitable for the actual environment and have been put into practice In to prove.