基于模糊逻辑推理的BP神经网络及其应用

Fuzzy logic reasoning-based BP neural network and its application

阐述了BP神经网络的原理及学习算法，在结合模糊逻辑推理的基础上提出了一种具有分层结构，能够进行规则自提取、自修正、自学习的复合模糊BP神经网络模型．这种模糊神经网络不仅可以充分利用原有的专家的经验和知识，而且能够从实际数据中通过不断学习获取新的知识和推理规则．同时，在相应的网络权值训练中引入了遗传算法和模糊逻辑控制器的优化求解思想．还进一步探讨了将这种网络模型用于汇率分析系统的形式和方法．

Introduces the principle and learning algorithm of BP neural network, and proposes a layered compound fuzzy BP neural network model capable of doing seif-extraction, seif-correction and self-leaming of rules, which not ouly makes full use of original expert experience and knowedge, but also acquires new knowedge and reasoning rules from actual data through continuous leaming, while introduces the inheritance algorithm and the concept of optimum solution by fuzzy logic controller into the crresponding network weighted value taining.