小波变换和RBF网络用于模式法分解重叠色谱峰

将小波变换和神经网络相结合，研究色谱数据处理中分解重叠峰的新方法.首先利用小波变换多分辨分析和奇异性检测原理提取重叠色谱峰上的各特征点，构造反映重叠峰形状、位置和高度的5个无因次特征量，然后借助径向基函数(RBF)网络来表达重叠峰中子峰面积比和5个无因次特征量的映射关系，建立分解重叠色谱峰的模式识别模型.实验结果表明，采用训练好的RBF神经网络分解重叠色谱峰，准确度优于传统的垂线分割法，而且可实现对只有一个峰强极大点的肩峰型重叠峰的分解.

Resolution of overlapped chromatographic peaks by pattern recognition based on wavelet transform and RBF networks

A novel method for resolving overlapped peaks with wavelet transform and neural network was proposed. The singularity detection based on multi-resolution analysis of wavelet transform was employed to extract characteristic points on overlapped chromatographic peaks. According to the positions and values of the characteristic points, five characteristic parameters characterizing the shape, position and height of individual peak in overlapped peaks were introduced. Radial basis function(RBF) neural networks were used to correlate the parameters with the area percentage of individual peak and build the model for recognizing the patterns of overlapped chromatographic peaks. Experimental results indicate that the proposed method has good performance with high accuracy and can resolve shoulder peaks in comparing with the conventional geometric method.