论文部分内容阅读
在D-FNN算法基础上,提出了基于椭圆基函数(EBF)的广义动态模糊神经网络。该算法提出模糊ε-完备性作为高斯函数宽度的确定准则,避免初始化选择的随机性;同时,该算法不仅能对模糊规则而且能对输入变量的重要性作出评价,从而使得每个输入变量和模糊规则都可以根据误差减少率(ERR)来修正。其应用不仅可以用来建模,还可以用来抽取有意义的模糊规则以获取知识。通过与D-FNN以及其他方法的比较,可以看到GD-FNN在学习效率和性能方面具有突出的优势。最后针对实际案例进行了仿真分析,验证了该算法的有效性和高效性。
Based on the D-FNN algorithm, a generalized dynamic fuzzy neural network based on elliptic basis function (EBF) is proposed. This algorithm proposed fuzzy ε-completeness as the criterion for determining the width of Gaussian function, and avoided the randomness of initial selection. At the same time, the algorithm not only evaluates the importance of fuzzy variables but also makes the input variables Fuzzy rules can be corrected according to the error reduction rate (ERR). Its application not only can be used to model, but also can be used to extract meaningful fuzzy rules to obtain knowledge. By comparing with D-FNN and other methods, we can see that GD-FNN has outstanding advantages in learning efficiency and performance. Finally, a simulation analysis is carried out on the actual case to verify the effectiveness and efficiency of the algorithm.