【作者】 任胜兵； 喻寿益；

【Author】 REN Shengbing YU Shouyi (Institute of Information Science & Engineering, Central South University, Changslta, 410083)

【机构】 中南大学信息科学与工程学院；

【摘要】 文章提出了一种正交基模糊神经网络(OBFNN),并给出了相应的构造方法,即正交模糊集法。该方法构造的正交基模糊神经网络参数物理含义明确,计算量少,且证明了构造得到的网络具有线性插值特性。仿真实验表明,此方法构造的正交基模糊神经网络对测试数据具有很高的拟合精度。更多还原

【Abstract】 This paper proposes a kind of orthogonal basis fuzzy-neural (OBFN) network , and discusses a construction method which is called orthogonal fuzzy sets (OFS) method The parameters of the OBFN network constructed by the OFS method have clear physical interpretation , and can be constructed through less calculation. It is proved that the OBFN network constructed by the OFS method has the linear interpolation characteristic. The experiments demonstrate that the OBFN network which is constructed by the OFS method can approximate the test data precisely.更多还原