摘要
通过相关分析,筛选出济阳坳陷控制成藏体系资源丰度的主要地质参数,把成藏体系的资源丰度及其主控地质参数作为神经网络的训练样本,建立了具有预测功能的神经网络结构。与多元线性和多元非线性回归模型预测结果的比较表明,神经网络模型在预测精度上要远高于2种回归预测模型。
The major geologic parameters for controlling the abundance of hydrocarbon resources of hydrocarbon accumulation in Jiyang Depression are selected by correlated analysis. The abundance of hydrocarbon resources and its major control geologic parameters are used as sample for neural network training, and structures of neural network with prediction functions are established. The result is contrasted with that of multi-element linear and multi-element nonlinear model prediction, it shows that the precision of neural network model is much higher than that of above 2 models.
出处
《石油天然气学报》
CAS
CSCD
北大核心
2007年第1期55-58,共4页
Journal of Oil and Gas Technology
关键词
神经网络
资源丰度
资源评价
济阳坳陷
neural network
abundance of resource
resource evaluation
Jiyang Depression
