论文部分内容阅读
数学可解性是利用神经网络算法求解问题需要首先考虑的问题，其次是用于训练的数据 集有效性．本文针对地层沉积相预测问题，从网络映射分析角度计算了三层神经网络容量能力，求 证了隐层节点与网络稳定性的关系，并给出内在关系式，从计算能力上分析神经网络用于岩性及 岩相预测的可行性，为克服神经网络映射的复杂性和训练数据的不确定性提供理论依据．
Mathematical solvability is a problem that needs to be considered first to solve the problem using neural network algorithm, and secondly, the effectiveness of the data set used for training. Aiming at the problem of stratum sedimentary facies prediction, this paper calculates the capacity of three-layer neural network from the perspective of network mapping and verifies the relationship between hidden layer nodes and network stability. The inner relationship is given, and the neural network is used to analyze The feasibility of lithology and lithofacies prediction provides a theoretical basis for overcoming the complexity of neural network mapping and the uncertainty of training data.