论文部分内容阅读
本文研究了径向基概率神经网络(Radial Basis Probabilstic Neural Networks,RBPNN)的一种新的无监督学习算法,该算法整合了径向基概率神经网络的结构原理与动态聚类算法的特点,使得在对训练样本的聚类分析并正确划分其类别属性的同时,自动完成径向基概率神经网络的训练过程.本算法在对IRIS和双螺旋分类问题的应用中,取得了较好的分类效果,而且在推广能力方面,由本文算法训练的RBPNN要明显好于有监督训练的径向基函数神经网络(RBFNN).
This paper studies a new unsupervised learning algorithm for Radial Basis Probabilistic Neural Networks (RBPNN), which combines the structural principle of radial basis probabilistic neural networks and the characteristics of dynamic clustering algorithms such that The training process of radial basis probabilistic neural network is completed automatically while clustering the training samples and classifying them correctly.The algorithm has achieved good classification results in the application of IRIS and double helix classification problems , And RBPNN trained by this algorithm is obviously better than RBFNN with supervised training in terms of promotion ability.