论文部分内容阅读
径向基函数网络构造中的关键问题是网络中心的选取,最小二乘算法采用正交化方法,独立计算回归算子对输出的贡献,故使中心的选择步骤简单有效。给出了最小二乘算法及其应用函数逼近的实例,结果证明,由于计算过程中应用了这一算法的正交化性质,所以网络调整时对已有模式的扰动达到最小。这说明最小二乘算法不仅简单有效,而且性能优越,并有较强的实用性,在许多领域有广泛应用。
The key problem in the construction of RBF network is the selection of network center. The least squares algorithm uses the orthogonalization method to calculate the contribution of regression operator to the output independently, so that the central selection step is simple and effective. The least square method and its application function approximation are given. The results show that the disturbance to the existing mode during network adjustment is minimized due to the orthogonal nature of this algorithm. This shows that the least squares algorithm is not only simple and effective, but also superior performance, and has strong practicality, has been widely used in many fields.