Growing radial basis neural networks: Merging supervised and unsupervised learning with network growth techniques
Growing radial basis neural networks: Merging supervised and unsupervised learning with network growth techniques
复制标题
DOI：
10.1109/72.641471
复制
发表时间：
1997-11-01
影响因子：
--
通讯作者：
Mi, GWQ
中科院分区：
文献类型：
--
作者：
Karayiannis, NB;Mi, GWQ
提出了一种构造和训练径向基函数（RBF）神经网络的框架。提出了一种生长型径向基函数网络（Growing Radial Basis Function，GRBF），该网络从确定径向基函数位置的少量原型开始，在训练过程中，GRBF网络通过在每个生长周期分裂一个原型来生长，提出了两个分裂准则来确定在每个生长周期分裂哪个原型，提出的混合学习方案提供了一个框架，将现有的算法在GRBF网络的训练，这些包括无监督算法的聚类和学习矢量量化，提出了一种基于局部类条件方差最小化的有监督学习方案，并对该方案进行了测试，在多种数据集上对GRBF神经网络进行了评估和测试，取得了令人满意的结果。
This paper proposes a framework for constructing and training radial basis function (RBF) neural networks. proposed growing radial basis function (GRBF) network begins with a small number of prototypes which determine the locations of radial basis functions, In the process of training, the GRBF network grows by splitting one of the prototypes at each growing cycle, Two splitting criteria are proposed to determine which prototype to split in each growing cycle, The proposed hybrid learning scheme provides a framework for incorporating existing algorithms in the training of GRBF networks, These include unsupervised algorithms for clustering and learning vector quantization, as well as learning algorithms for training single-layer Linear neural networks, A supervised learning scheme based the minimization of the localized class-conditional variance also proposed and tested, GRBF neural networks are evaluated and tested an a variety of data sets,vith very satisfactory results.