二进小波神经网络用于示波计时电位测定

提出了二进小波神经网络的结构及算法，并用于单组分和多组分示波计时电位信号的浓度计算。在二进小波神经网络中选用了Morlet母小波和修理的误差反传前向神经网络。探讨了二进小波神经网络中的中小波基个数，初始学习速率因子和动量因子等参数对网络预测结果的影响。结果表明：二进小波神经网络对双组分和单组分示波计时电位信号中去极剂浓度的预测均有很好效果。     

一种基于二进小波变换的自适应滤波方法

根据信号和噪声经小波变换后在不同尺度上有不同的特征，将相邻尺度二进小波变换值的相关量进行归一化处理并与小波变换值比较来判断信号与噪声，以噪声在各尺度的方差作为终止迭代的标准，提出了一种基于二进小波变换小波域选择噪声的自适应滤波方法。研究了模拟信号的去噪过程、半峰宽和信噪比对去噪结果的影响，并对模拟含噪信号和含噪毛细管电泳信号去噪前后的结果进行了比较。实验结果表明：由于该方法具有良好的自适应性和显著的滤波效果，必将得到广泛的应用。     

小波变换用于色谱重叠峰的解析

利用小波变换的时频局部化性质，通过对色谱重叠峰信号小波变换后的某些频率段进行放大，使重叠谱峰得到了分离，并将此方法用于苯和甲苯二组分色谱体系的定量分析，重叠峰中各组分均得到了良好的线性关系及令人满意的定量分析。本文还讨论了不同小波基、分解的次数及放大系数在解析结果的影响。     

小波变换在差分脉冲伏安法数据处理中的应用

本文根据小波变换不改变信号线性关系的性质，用ＤＯＧ小波函数处理了Ｃｕ－ＫＮＯ３中Ｃｕ＾２＋的差分脉冲伏安法实验数据，结果延长了测定的线性范围。     

连续小波变换-独立成分回归算法及其在多组分分析中的应用

采用连续小波变换（CWT）对光谱数据进行处理，用独立成分分析（ICA）进行特征提取，再用回归分析方法对被测组分进行测定，建立了连续小波变换一独立成分回归（CWT-ICR）方法。方法用于肉样品中水分、脂肪和蛋白质多组分的同时测定，所得结果与化学法测得结果相符。     

分析化学计量学中的新方法─—小波分析

小波分析在分析信号处理中具有诸多显著的优点。本文介绍了小波分析的一般描述，综述了化学计量学在的小波新方法，并展望了小波分析在分析化学计量学中的应用前景。     

分析化学计量学中的新方法——小波分析

小波分析在分析信号处理中具有诸多显著的优点。本文介绍了小波分析的一般描述，综述了化学计量学在的小波新方法，并展望了小波分析在分析化学计量学中的应用前景。     

小波神经网络用于钼和钨的同时测定

介绍了小波神经网络的结构和算法，并首次用于混合试样中的钼和钨含量测定，在小波神经网络中，采用Ｍｏｒｌｅｔ母小波和一维搜索变步长共轭梯度优化方法，结果表明，小波神经网络优于ＢＰ网络，其预测含量的回收率在９６．０％～１０４．６％之间。     

用墨西哥帽小波研究DNA序列的分形特征

结合小波分析和分形理论，采用分数布朗运动(FBM)建立数学模型，研究脱氧 核糖核酸(DNA)序列的自相似性。用DNA walk的方式将DNA序列表达成为一个数字信 号，通过不同小波变换的尺度对应不同特征长度的碱基，选择墨西哥帽小波为母小 波，进行实验考察，结果发现小波系数的图形在许多尺度看上去很相似，大尺度对 应较多的碱基(小波变换尺度为2^7时，对应512个碱基)，能看到概貌；小尺度对应 较少的碱基(小波变换尺度为2^3时，对应32个碱基)，可看到细节。这表明其DNA序 列中存在分形结构，可以用分维数来作为定量描述。这种算法为进一步研究与基因 序列自相似结构有关的基因进化信息提供一种选择的途径。     

毛细管电泳激光诱导荧光检测信号的小波滤噪

采用小波滤噪方法对毛细管电泳激光诱导荧光检测信号进行了处理，研究了小波变换中小波基的选择及噪声阈值的选择对滤噪的影响。结果表明，采用DB4小波基能有效消除毛细管电泳激光诱导荧光检测信号中存在的噪声，使信噪比得到较大改善。     

一种基于小波变换的近红外化学指纹图谱分析方法

提出了一种快速无损的化学指纹图谱分析新方法. 根据小波变换多分辨分析的特点, 对近红外(NIR)光谱进行小波分解, 从中提取被测样品的化学特征信息, 并作数字信息可视化处理, 构建形成可直观识别样品模式特征的NIR指纹图谱. 将该方法用于中药材丹参的质量检测, 检测结果与色谱指纹图谱检测结果相符, 能快速有效地识别丹参质量模式间的差异, 有望发展成为一种快速分析检测天然产物质量的方法.     

小波变换用于高效液相色谱的噪声滤除

根据小波变换将信号分频的性质，用Ｈａａｒ小波成功地解决了高效液相色谱中噪音的扣除问题，并应用于乳酸－稀土络合物体系的色谱数据，得到的色谱曲线平滑且峰位置不变，大葱旁氡群徒档土思觳庀蕖?     

The application of digital image recognition to the analysis of two-dimensional fingerprints

High-performance liquid chromatography (HPLC) fingerprint has been commonly used in the quality control and assessment of herbal medicines, and two-dimensional (2D) fingerprint obtained by means of HPLC-diode array detector (HPLC/DAD) can provide higher reliability. In this paper, an approach to the analysis of the 2D HPLC/DAD fingerprints, which was based on digital image recognition techniques, was developed for the first time. First, wavelet transform was employed to eliminate noise signal in the 2D fingerprint, and then the 2D fingerprint was converted into grayscale image. Second, several features of the image were calculated, and hierarchical clustering. This approach was applied to the qualitative analysis of the different samples of coptis chinensis, and the clustering result of samples was all highly consistent with the real situation. Based on the densities in grayscale image, three components in standard samples were quantitative analyzed, and the obtained correlation coefficients between concentration and grayscale density were more than 0.999. Our study indicated that the analysis of the 2D HPLC/DAD fingerprint was successful based on the idea and techniques of digital image recognition techniques, and this proposed approach provided a new pathway for the analysis of two-dimensional spectrums.     

Maximum spectrum of continuous wavelet transform and its application in resolving an overlapped signal

To estimate the number of peaks and to find the individual peak positions in an overlapped signal, a new method called maximum spectrum of continuous wavelet transform (MSCWT) was developed by extracting the maximum coefficients of continuous wavelet transform (CWT). The peak position in MSCWT was the same as that in its original signal. In this process, CWT was performed not on a single dilation but on an appreciation dilation range. To obtain such a range, a new criterion was introduced to choose a center dilation, which was used to form the dilation range. If Cdilation denoted the center dilation, the proper dilation range was [Cdilation -6 +/- 2, Cdilation +1 +/- 1]. The Mexican Hat function was an analytical wavelet. Utilizing the information of the peak number and the position detected by MSCWT, a fitting route was performed to recover the original signal. One simulated and four true overlapped signals, including high performance liquid chromatography (HPLC), ultraviolet-visible (UV) spectrum, and differential pulse voltammetric (DPV), were processed, and the results indicated that MSCWT could detect an overlapped peak number and position, and the curve fitting based on information of MSCWT had a higher accuracy. The proposed method was an efficient one in resolving different types of overlapped signals.     

Improving detection power in trace analysis using wavelet transform

Environmental analysis most often is trace analysis. Therefore, the concentrations are commonly in the lower working range near the limit of detection of the corresponding analytical method. However, whenever the instrument's analytical noise is too large, it dominates the signal curves and analytes cannot be detected anymore. Furthermore, the evaluation of peaks with defined baselines is hindered very much. One possibility for de-noising is wavelet transform which is presented in this work. Different wavelet functions are applied and Symlet4 is suggested as the most powerful for analytical peaks that resemble Gaussian distribution curves, as it improves limits of detection by factors 6 to 7. The comparison of different wavelet functions has been carried out for two modern analytical scopes. At first, chromatograms are de-noised for the speciation of four arsenic compounds via the coupling of HPLC and ICP-MS. Secondly, the determination of cadmium is shown by HR-CS AAS, which is one of the most recently developed devices in atomic absorption spectrometry and allows the registration of three-dimensional spectra in order to investigate the spectral vicinity of analytical lines. On the basis of these investigations, we recommend using wavelet transform with Symlet4 for all analytical techniques which are resulting in similar signal curves.     

小波变换与分析化学信号处理

介绍了小波变换的基本理论并对小波变换的常用算法和应用进行了评述。由于小波变换的时2频局部化性质, 使其成为信号处理的强有力工具。在分析化学领域中, 小波变换在流动注射分析、伏安分析、高效液相色谱、红外光谱、质谱、核磁共振谱、可见-紫外光谱、光声光谱、扩展X-射线吸收精细结构(EXAFS) 谱等分析化学信号的平滑滤噪、数据压缩、重叠信号解析等方面都有成功的应用。     

基于小波神经网络的新型算法用于化学信号处理

基于紧支集正交小波神经网络的构造思想,用具有紧支集的Ｂ－样条函数的伸缩和平稳替代小波函数,提出了一种新型算法,并将其应用于化学信号的处理,实现了信号的压缩和滤噪,适应小波神经网络相比,学习速度得到了大幅度的提高。     

Comparison of spline wavelet and orthogonal wavelet in processing analytical chemistry signals

Spline wavelet and orthogonal wavelet are two widely used wavelet methods. In this paper, comparison of these two methodshas been made, including their algorithm, properties and results of signal processing in analytical chemistry signals. It is found that spline wavelet is more effective than orthogonal wavelet in processing high noise signals. The curves obtained from spline wavelet are closer to the theoretical ones than those obtained from orthogonal wavelet and the errors of spline wavelet are smaller than those of orthogonal wavelet.     

Determination of Total Sugar in Tobacco by Near-Infrared Spectroscopy and Wavelet Transformation-Based Calibration

Based on wavelet transformation (WT) and mutual information (MI), a simple and effective procedure is proposed for multivariate calibration of near-infrared spectroscopy. In such a procedure, the original spectra of the training set are first transformed into a set of wavelet representations by wavelet prism transform. Then, the MI value between each wavelet coefficient variable and the dependent variable is calculated, resulting in a MI spectrum; by retaining a subset set of coefficients with higher MI, an update training set consisting of wavelet coefficients is obtained and reconstructed/converted back to the original domain. Based on this, a partial least square (PLS) model can be constructed and optimized. The optimal wavelet and decomposition level are determined by experiment. A NIR quantitative problem involving the determination of total sugar in tobacco is used to demonstrate the overall performance of the proposed procedure, named RPLS, meaning PLS in reconstructed original domain coupled with MI-induced variable selection in wavelet domain (RPLS). Three kinds of procedures, that is, conventional full-spectrum PLS in original domain (FPLS), PLS in original domain coupled with MI-induced variable selection (OPLS), and direct PLS in MI-based wavelet coefficients (WPLS), are used as reference. The result confirms that it can build more accurate and robust calibration models without increasing the complexity.     

Wavelet multiscale regression from the perspective of data fusion: new conceptual approaches

Wavelet regression is a very promising technique for modern multivariate calibration and calibration transfer. Multiscale analysis of wavelet scales provides a connection between wavelet regression and data fusion. In this paper, current wavelet regression methods are reviewed from the novel perspective of data fusion. Illustrated by analysis of a public domain near-infrared dataset, the advantages and drawbacks of these methods are examined. For wavelet regression, the non-uniformity of the wavelet components, the multiscale nature of the signal, and the prevention of information leakage are crucial issues that will be addressed.