论文部分内容阅读
针对传统人工神经网络在解决时间序列数据挖掘问题时受到输入同步瞬时限制的问题,本文提出了时序过程神经元网络挖掘模型,并给出了基于离散Walsh函数变换的学习算法。模型在处理时间序列数据挖掘问题时,能够充分反映时间序列中实际存在的时间累积效应,其精度和泛化能力都高于传统人工神经元网络。本文最后将模型应用于油田地质领域中的水淹层识别的实际问题中,应用实例证明了模型和算法的有效性。
Aiming at the problem that the traditional artificial neural network is limited by the instantaneous input synchronization while solving the problem of time series data mining, this paper proposes a time series neural network mining model and a learning algorithm based on the discrete Walsh function transformation. When dealing with the problem of time series data mining, the model can fully reflect the actual time accumulation effect in time series, and its accuracy and generalization ability are higher than the traditional artificial neural network. In the end of this paper, the model is applied to the practical problems of water flooded layer recognition in the field of oilfields. The application examples are given to demonstrate the effectiveness of the model and the algorithm.