摘要
小波分析是一种新的信号处理技术 ,具有良好的时频局部化特征。为了克服BP网络自身算法的缺陷 ,得到更高的学习精度和更快的收敛速度 ,使用小波包分析的特征提取及神经网络的非线性映射特性 ,构造了小波神经网络 ,以此为基础开发出的软件系统具有使用的特征量少 ,建造预报系统较为简单等优点。将之应用于炼铜转炉炉渣重量及成分预报 ,该模型完全能够较准确地预报出渣量和成分。其平均拟合误差为 1 5 % ,平均预报误差为 3 1%
Wavelet analysis,a new technique for signal processing,possesses excellent characteristic of time frequency localization, and is suitable for analyzing the time varying or transient signals In order to overcome the algorithm shortcoming of BP network and to obtain much higher accuracy and faster speed,a wavelet neural network is put forward by means of feature extraction of wavelet package analysis and nonlinear mapping of neural network Based on this model,a software system is developed with fewer characteristic quantity and being built up easily The software is applied to forecast slag weight and components Results showed that the model could forecast the slag weight and components accurately The average error is 1 5 percent and the average forecasting error is 3 1 percent
出处
《有色金属》
CSCD
2001年第2期42-44,共3页
Nonferrous Metals
关键词
小波分析
神经网络
炼铜
砖炉
炉渣
成分预报
wavelet analysis
neural network
copper smelting converter
slag
forecasting