学习向量量化神经网络用于胃癌组织样品分类识别的研究

将lvq神经网络(Learn ing Vector Quantization Neural Networks)用于胃癌组织样品的分类识别,根据胃癌组织及相应正常组织的FTIR光谱的主要特征吸收峰值(包括vas(CH3)、vs(CH2)、δ(CH2)、v(C-O)、vs(PO2-)、vas(PO2-)和vs(核酸,细胞蛋白及膜脂))全部或部分作为网络输入向量,对未知的胃组织样品进行分类识别,结果显示:i)以上述全部七个谱峰为输入向量时,网络经训练学习后,其平均识别正确率最高(达89.3%),表明该网络对胃癌组织样品的分类识别是满意的,完全可作为临床医学的辅助诊断手段;ii)总体上,当作为输入向量的FTIR特征谱峰越多时,则网络的平均分类识别正确率越高;iii)作为输入的FTIR特征谱峰不同时,则网络的平均分类识别正确率也不同。     

Selection of quasi-optimal inputs in chemometrics modeling by artificial neural network analysis

Instrumentation spectra used for chemometrics analysis are often too unwieldy to model, as many of the inputs do not contain important information. Several mathematical methods are used for reducing the number of inputs to the significant ones only. Artificial neural networks (ANN) modeling suffers from difficulties in training models with a large number of inputs. However, using a non-random initial connection weight algorithm and local minima avoidance and escape techniques can overcome these difficulties. Once the ANN model is trained, the analysis of its connection weights can easily identify the more relevant inputs. Repeating the process of training the ANN model with the reduced input set and the selection of the more relevant inputs can proceed until a quasi-optimal, small, set of inputs is identified. Two examples are presented—finding the minimal set of wavelengths in benchmark diesel fuel NIR spectra, and in spectra generated in a recent work, modeling of “artificial nose” sensor array. In the last example, 1260 inputs were reduced to optimal sets of <10 inputs. Causal index calculation can analyze the influence of each of selected wavelengths on the predicted property. Some of the resulting minimal sets are not unique, depending on the ANN architecture used in the training. The accuracy of the resulting ANN models is usually better, and more robust, than the original large ANN model.     

色谱重叠峰分解的神经网络法

对于色谱图中两个相互重叠的峰, 本文提出了一种分解方法。首先在重叠峰的一阶导数曲线上取出五个无因次特征值, 然后用多层前传网来表达这五个特征值和重叠峰中子峰面积分率之间的关系。一系列实验的结果表明, 用神经网络方法所得子峰面积的准确度, 优于传统的垂线法及函数拟合法, 而且计算工作量较小, 可用于实时处理。     

人工神经网络法同时测定苯二酚异构体

研究了人工神经网络法与紫外分光光度分析相结合，用于邻苯二酚、间苯二酚、对苯二酚异构体的含量测定。本研究采用人工神经网络法，直接对苯二酚异构体混合液的紫外吸收光谱数据进行预测，不需预先分离，即可得到各异构体的浓度。     

Dynamic modelling of milk ultrafiltration by artificial neural network

Artificial neural networks (ANNs) have been used to dynamically model crossflow ultrafiltration of milk. It aims to predict permeate flux, total hydraulic resistance and the milk components rejection (protein, fat, lactose, ash and total solids) as a function of transmembrane pressure and processing time. Dynamic modelling of ultrafiltration performance of colloidal systems (such as milk) is very important for designing of a new process and better understanding of the present process. Such processes show complex non-linear behaviour due to unknown interactions between compounds of a colloidal system, thus the theoretical approaches were not being able to successfully model the process. In this work, emphasis has been focused on intelligent selection of training data, using few training data points and small network. Also it has been tried to test the ANN ability to predict new data that may not be originally available. Two neural network models were constructed to predict the flux/total resistance and rejection during ultrafiltration of milk. The results showed that there is an excellent agreement between the validation data (not used in training) and modelled data, with average errors less than 1%. Also the trained networks are able to accurately capture the non-linear dynamics of milk ultrafiltration even for a new condition that has not been used in the training process.     

Prediction of protein secondary structure content by artificial neural network

The neural network method was applied to the prediction of the content of protein secondary structure elements, including alpha-helix, beta-strand, beta-bridge, 3(10)-helix, pi-helix, H-bonded turn, bend, and random coil. The "pair-coupled amino acid composition" originally proposed by K. C. Chou [J Protein Chem 1999, 18, 473] was adopted as the input. Self-consistency and independent-dataset tests were used to appraise the performance of the neural network. Results of both tests indicated high performance of the method.     

Application of artificial neural network in food classification

Artificial neural network (ANN) classifiers have been successfully implemented for various quality inspection and grading tasks of diverse food products. ANN are very good pattern classifiers because of their ability to learn patterns that are not linearly separable and concepts dealing with uncertainty, noise and random events. In this research, the ANN was used to build the classification model based on the relevant features of beer. Samples of the same brand of beer but with varying manufacturing dates, originating from miscellaneous manufacturing lots, have been represented in the multidimensional space by data vectors, which was an assembly of 12 features (% of alcohol, pH, % of CO(2) etc.). The classification has been performed for two subsets, the first that included samples of good quality beer and the other containing samples of unsatisfactory quality. ANN techniques allowed the discrimination between qualities of beer samples with up to 100% of correct classifications.     

A comparative study of diesel analysis by FTIR, FTNIR and FT-Raman spectroscopy using PLS and artificial neural network analysis

Diesel properties determined by ASTM reference methods as cetane index, density, viscosity, distillation temperatures at 50% (T50) and 85% (T85) recovery, and the total sulfur content (%, w/w) were modeled by FTIR-ATR, FTNIR, and FT-Raman spectroscopy using partial last square regression (PLS) and artificial neural network (ANN) spectral analysis. In the PLS models, 45 diesel samples were used in the training group and the other 45 samples were used in the validation. In the ANN analysis a modular feedforward network was used. Sixty diesel samples were used in the neural network training and other 30 samples were used in the validation. Two different ATR configurations were compared in the FTIR, a conventional (ATR1) and an immersion (ATR2) cell. The ATR1 cell presented the best results, with smaller prediction errors (root mean square error of prediction, RMSEP). The comparison of the three PLS models (FTIR-ATR1, FTNIR, and FT-Raman) shows that reasonable values of R² and RMSEP were obtained by the FTIR-ATR1 and FTNIR models in the evaluation of density, viscosity, and T50. The PLS/FT-Raman models presented reasonable results only for the T50 property. None of the techniques was able to generate suitable PLS calibration models for the determination of sulfur content. The ANN/FT-Raman models presented the best performances, with all models presenting R²-values above 85% some of them with RMSEP values significantly smaller than those obtained with FTIR-ATR and FTNIR. The ANN/FT-Raman and ANN/FTIR-ATR1 models were able to estimate the total sulfur content of diesel with 0.01% (w/w) accuracy.     

Recognition of adulteration of Italian wines by thin-film multisensor array and artificial neural networks

The paper reports the use of a chemoresistive multisensor array for recognition of some adulterated Italian wines (two white, four red, two rosè) added with methanol, ethanol or other same-colour wine. A multisensor array constituted by four thin-film semiconducting metal oxide sensors, surface-activated by Pt, Au, Pd, Bi metal catalysts, has been used to generate the chemical pattern of the volatile compounds present in the wine samples. The responses of the multisensor array towards wines tested by headspace sampling have been evaluated. Multivariate analysis including principal component analysis (PCA) as well as back-propagation method trained artificial neural networks (ANNs) have been applied to analytical data generated from the multisensor array to identify both the adulteration of wines and to determine the added content of adulterant agent of methanol or ethanol up to 10 vol.%. The cross-validated ANNs provide the highest achieved percentage of correct classification of 93% and the highest achieved correlation coefficient of 0.997 and 0.921 for predicted-versus-true concentration of methanol and ethanol adulterant agent, respectively.     

Systematic optimization of the analysis of wine bouquet components by solid-phase microextraction

The extraction, identification, and quantification of wine aroma compounds are preliminary steps required for further investigation of wine quality, i.e. determination of the varieties of grapes used, the production process, and the origin and age of the wine. This paper deals with the optimization of solid-phase microextraction for the determination of compounds which produce wine bouquet. Optimum operating conditions have been determined to obtain high reproducibility at low cost and with low time-consumption. Several factors influencing the equilibrium of the aroma compounds between the sample and the fiber must be taken into account, including length of contact time between the two phases involved, speed of agitation of the sample, the matrix in which the process takes place, and, furthermore, the place, duration, and temperature of desorption in the injector of the chromatograph.     

人工神经网络用于光度法测定工业废水中的铬和铁

为了快速测定工业废水中的铬和铁,利用BP人工神经网络模型,并与分光光度法相结合,不经分离同时测定工业废水中的铬和铁。在铬-铁-二苯碳酰二肼显色体系中,控制pH在1.0-2.0之间,用分光光度法测定显色体系的吸收光谱,应用三层人工神经网络解析吸收光谱,同时测得铬和铁的浓度。详细研究了分光光度法同时测定铬和铁的测定条件和网络训练的最佳训练参数,BP人工神经网络的动量参数为0.8,拓扑结构为20-15-2,转换函数的形状参数为0.9。该测定方法不仅可用于环境监测,而且能用于食品、材料、药物、生物样品、矿物等物质中铬和铁的测定。     

Application of quantitative artificial neural network analysis to 2D NMR spectra of hydrocarbon mixtures

Understanding relationships between the structure and composition of molecular mixtures and their chemical properties is a main industrial aim. One central field of research is oil chemistry where the key question is how the molecular characteristics of composite hydrocarbon mixtures can be associated with the macroscopic properties of the oil products. Apparently these relationships are complex and often nonlinear and therefore call for advanced spectroscopic techniques. An informative and an increasingly used approach is two-dimensional nuclear magnetic resonance (2D NMR) spectroscopy. In the case of composite hydrocarbons the application of 2D NMR methodologies in a quantitative manner pose many technical difficulties, and, in any case, the resulting spectra contain many overlapping resonances that challenge the analytical work. Here, we present a general methodology, based on quantitative artificial neural network (ANN) analysis, to resolve overlapping information in 2D NMR spectra and to simultaneously assess the relative importance of multiple spectral variables on the sample properties. The results in a set of 2D NMR spectra of oil samples illustrate, first, that use of ANN analysis for quantitative purposes is feasible also in 2D and, second, that this methodology offers an intrinsic opportunity to assess the complex and nonlinear relationships between the molecular composition and sample properties. The presented ANN methodology is not limited to the analysis of NMR spectra but can also be applied in a manner similar to other (multidimensional) spectroscopic data.     

Predicting chain dimensions from an artificial neural network model

Artificial neural network models are used to investigate polymer chain dimensions. In our model, the input nodes are glass transition temperature (T_g), entanglement molecular weight (M_e), and melt density (ρ). The number of nodes in the hidden layer is eight. We found that the relative error for prediction of the characteristic ratio ranges from 0.77 to 7.5% and that the overall average error is 3.57%. Artificial neural network models may provide a new method for studying statistics properties of polymer chains. © 2000 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 38: 3163–3167, 2000     

Analysis of ultrasonic velocity in refrigerants using artificial neural network

Several researchers have reported numerous measurements on ultrasonic velocity as a function of temperature and pressure using various experimental techniques. A large amount of experimental data is required in order to obtain accurate results for the chemical substances used. The present article explores the evaluation of ultrasonic velocity as a function of molecular weight, temperature and pressure using an artificial neural network (ANN) in six refrigerants. The network so developed predicts the ultrasonic velocity successfully. Statistical analysis of the results was performed using standard deviation (%) and relative average deviation. The correlation coefficient in our analysis was found to be 0.9999. The trained weights, obtained from ANN, are further employed to form equations to predict ultrasonic velocity at other temperatures and pressures.