论文部分内容阅读
基于对目前神经网络存在问题的具体分析，认为将启发性信息引入神经网络训练将是提高网络学习能力＼质量以及效率的重要途径。进而讨论了启发知识的来源与种类，将启发性知识分成诱导性约束和强制性约束两类，进而建立了引入网络训练的相应策略，给出了启发性知识引入与选择的具体原则，并建立了两种基于导数关系的启发知识模型。最后建立了神经网络的具体训练算法。具体应用结果证明了所提出策略与方法的有效性。
Based on the specific analysis of the existing problems in the neural network, it is considered that introducing the heuristic information into the neural network training will be an important way to improve the network learning ability, quality and efficiency. Furthermore, the sources and categories of enlightenment knowledge are discussed, and the enlightenment knowledge is divided into two categories: inductive constraint and mandatory constraint, and then the corresponding strategies of introducing network training are established. The specific principles of enlightenment knowledge introduction and selection are also given Two kinds of heuristic knowledge model based on the derivative relationship. Finally, the specific training algorithm of neural network is established. The application results prove the effectiveness of the proposed strategy and method.