摘要
介绍了小波神经网络（ＷａｖｅｌｅｔＮｅｕｒａｌＮｅｔｗｏｒｋ）的结构和算法，并将其应用于色谱数据的滤噪与压缩。在小波神经网络中，采用了Ｂｕｂｂｌｅ母小波和一维搜索变步长共轭梯度优化方法。结果表明，小波神经网络将原始数据压缩至１∶２０时，仍能得到很好的重建谱图。重建的色谱图不仅能保留原始的谱图特征而且有光滑滤噪的效果。对于色谱数据的存储与平滑滤噪具有重要意义。
Based on wavelet analysis ,a new model of artificial neural network wavelet neural network(WNN) was proposed and applied to the compression and de noising of chromatograms.With the basic features of the chromatograms being well reserved,the original data of chromatograms can be compressed and represented using the WNN parameters.The compression ratio of 1∶20 was achieved.To confirm the effectiveness of this method in de noising,the reconstructed chromatograms at different training stages were presented and compared,which showed that the noise contained in the original chromatograms could be cleanly eliminated after compression.It was also found that the method is fast in convergence speed and accurate in reconstruction.
出处
《分析试验室》
CAS
CSCD
1998年第6期1-4,共4页
Chinese Journal of Analysis Laboratory
基金
国家自然科学基金
关键词
小波神经网络
色谱
数据压缩
滤噪
Wavelet neural network
chromatogram
de noising
compression
