摘要
鉴于BP网络训练时间过长,且易于陷入局部最优解,本文采用RBF网络来实现元音字母的语音识别。RBF网络的构造通过一种动态自适应聚类算法来完成,使得RBF网络具有在线学习能力。示例计算结果表明,这种RBF网络具有比BP网络和贝叶斯分类器更好的分类精度。
BP network needs too much time to train and probably plunges into local optimization. A RBF network is designed to implement phonetic classification of the vowel character. The RBF network is designed by a dynamic adaptive clustering algorithm and the network has the ability of on-line learning. The test result indicates that the classification precision of the RBF network is better than that of BP network and Bayesian classifier.
出处
《计算机与现代化》
2006年第7期15-17,共3页
Computer and Modernization
关键词
径向基函数
神经网络
元音识别
radial basis function
neural network
vowel classification
