摘要
针对BP神经网络收敛速度慢、容易陷入局部极小的缺陷,将遗传算法全局寻优的特点与BP神经网络相结合,利用遗传算法对神经网络的权值和阈值进行优化,构成一个GABP神经网络,有效地解决了BP神经网络容易陷入局部最优的问题。实验结果表明:相对于BP神经网络的说话人识别系统,基于GA-BP神经网络的说话人识别系统具有更快的网络收敛速度和更高的系统识别率。
Aiming at slow convergence and being easy to fall into local minimum problems in BP neural network,and combining with the global optimization future of genetic algorithm to improve weight and threshold value,a GA-BP neural network model is established.The results show that the GA-BP neural network-based speaker recognition system has a faster network convergence speed and higher recognition rate compared with BP neural network.
出处
《重庆理工大学学报(自然科学)》
CAS
2013年第10期91-95,共5页
Journal of Chongqing University of Technology:Natural Science
基金
福建省科技厅重点项目(2013H0002)
关键词
说话人识别
BP神经网络
遗传算法
权值
speaker recognition
GA-BP neural network
genetic algorithm
weight