摘要
针对标准的BP神经网络对于声音信号在线监控模型的预测误差比较大,提出了一种用遗传算法优化BP神经网络的算法,建立了声音监控的预测模型。遗传算法优化BP神经网络主要是用遗传算法来优化BP神经网络的初始权值和阀值,然后通过训练BP神经网络以得到预测模型的最优解,优化后的神经网络具有预测误差比较小、反应速度快等特点。实验结果证明,利用遗传算法优化BP神经网络在声音的智能监控中取得了比较好的效果,达到了系统设计的目的。
According to the standard BP neural network for relatively large prediction error in sound signal online monitoring model,a kind of algorithm using genetic algorithm to optimize BP neural network is put forward,and a sound monitoring prediction model is established.BP neural network optimized by genetic algorithm mainly uses the genetic algorithm to optimize BP neural network′s initial weights and threshold value,the optimal solution of prediction model is obtained through training BP neural network and the optimized neural network has a relatively small prediction error and a relatively fast speed,etc.The experimental result shows that a better effect is obtained by adopting BP neural network optimized by the genetic algorithm in intelligent sound monitoring,and the aim of the system design is achieved.
出处
《常州大学学报（自然科学版）》
CAS
2012年第3期70-74,共5页
Journal of Changzhou University：Natural Science Edition
基金
浙江省科技厅公益性项目资助(2011C31045)
关键词
智能监控
遗传算法
BP神经网络
信号处理
intelligent monitoring
genetic algorithm
BP neural network
signal processing
