RBF神经网络在视频业务建模及预测中的应用  

An application of RBF neural network in video traffic modeling and predicting

作　　者：熊思民 杨志民[1] 雷振明[1] 

机构地区：[1]北京邮电大学ATM技术研究中心,北京100876

出　　处：《通信学报》2000年第4期18-24,共7页Journal on Communications

基　　金：重庆市科委和原邮电部青年科技基金

摘　　要：ATM技术的关键在于业务控制 ,能否有效地实施业务控制则取决于对业务特征的了解和预测能力。传统的解析方法对视频业务进行预测的局限性较为明显。本文采用径向基函数神经网络对视频业务进行建模和预测 ,并提出了分别采用LBG算法和Hestenes奇异值分解进行隐含层神经元中心选择和输出神经腱权值计算的改进方法。文中利用一组实际的视频序列对该方法的有效性进行了验证。One of the key technologies of ATM is traffic control Whether a traffic control solution is effective depending on the through understanding of the characteristics and the ability of prediction The traditional analytical methods are obviously intractable for video traffic As a settlement,a radial basis function(RBF)neural networks has been adopted in this paper for video traffic modeling and prediction In addition,a proposal concerning the LBG basis algorithm and Hestenes singular value decomposition(SVD)method are involved,to be able to improve the selection of the centers of hidden layer neuron and the calculation method of the weights of the output layer synapse in the RBF network A series of practical video data has been used to evaluate the proposed approach

关 键 词：视频业务模型 径向基函数 预测 RBF神经网络 

分 类 号：TN919.8[电子电信—通信与信息系统]