摘要:

AbstractIn applications of pattern recognition techniques to problems in chemical fingerprinting, only limited knowledge about the underlying statistical distribution of the data is generally available. This means that non‐parametric methods must be used. Non‐parametric discriminant functions have been used to provide insight into relationships contained within sets of chemical measurements. However, classification based on random or chance separation can be a serious problem. Monte Carlo simulation studies have been carried out to assess the probability of chance classification for non‐parametric linear discriminants. The level of expected chance classification is a function of the number of observations (the number of samples), the dimensionality of the problem (the number of independent variables per observation), class membership distribution and the covariance structure of the data being examined. An approach for assessing the level of significance of classification scores obtained from real training sets will be presented. These simulation studies establish limits on the approaches that can be taken with real data sets so that chance classification are improbable, and provide information necessary for integrating the data analysis into the overall experimental d

ISSN:0886-9383    

DOI:10.1002/cem.1180020103

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractBased on a statistical model of the reproducibility of NMR spectral features, a system for computer retrieval of high‐resolution1H‐NMR spectra of glycoprotein carbohydrates has been developed. For corresponding peaks in an unknown and a reference spectrum, a similarity index based on the reproducibility of the chemical shifts is calculated. In addition, a second similarity index, based on the probability distribution of the percentage of non‐matching peaks, has been developed. From these two similarity indices, a combined similarity index using the recall–reliability function as the optimizing criterion has been derived.First results indicate that the ‘1H‐NMR reproducibility‐based retrieval’ (‘1HRR’) system offers good perspectives for both identification and su

ISSN:0886-9383    

DOI:10.1002/cem.1180020104

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractGas chromatography and pattern recognition methods have been used to develop a potential method for differentiating between European and Africanized honey‐bees based on chemical constitution. 243 European, African and Africanized honey‐bees were characterized by 40‐peak GCs of cuticular hydrocarbon extracts. Discriminants were developed that correctly classified the bees, and these discriminants were used successfully to classify bees of unknown origin, including hy

ISSN:0886-9383    

DOI:10.1002/cem.1180020105

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractA general problem arising in the development of regression models is the selection of the optimal model. Whenever a feature selection procedure, such as step forward, backward elimination, best subset or all possible combinations, or when a data compression approach, such as principal components or partial least‐squares regression, is used, the question of how many regression terms to include in the final model must be addressed.This work describes the evaluation of four different criteria for selection of the optimal predictive regression model using cross‐validation. The results obtained in this work illustrate the problems which can arise in the analysis of small or inadequately sampled data sets. The common approach, selecting the model which yields the absolute minimum in the predictive residual error sum of squares (PRESS), was found to have particularly poor statistical properties. A very simple change to a criterion based on the first local minimum in PRESS will provide a significant improvement in the cross‐validation result. A criterion based on testing the significance of incremental changes in PRESS with anF‐test may provide more robust performance than the local minimum in PRESS

ISSN:0886-9383    

DOI:10.1002/cem.1180020106

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractAn expert system (LABGEN) is presented for decision support in analytical laboratories by means of digital simulation. In an interactive manner, LABGEN constructs simulation models of laboratory organizations. It makes use of a database of model fragments and applies rules in order to prevent the user providing redundant information and to prevent inconsistent models being constructed. The models are written in the dedicated simulation language SIMULA. After compilation they can be used for simulation experiments. LABGEN can be applied to a wide range of laboratory organizations. An example of the application of LABGEN is presented.

ISSN:0886-9383    

DOI:10.1002/cem.1180020107

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractMethods of multivariate calibration use models that relate spectral data or sensor array responses to the concentrations of analytes. The goal is to insure that the calibration model can accurately estimate analyte concentrations in unknown samples not contained in the calibration set. The sensors or spectral channels (e.g. wavelengths) selected for incorporation in the model, as well as the samples selected for the calibration step, are known to have an effect on the accuracy of analysis for unknown samples. This work provides a fundamental treatment of this effect and derives criteria for optimal selection. Additionally, a proof is given for the advantage of having more sensors and calibration samples than analytes—the overdetermined cas

ISSN:0886-9383    

DOI:10.1002/cem.1180020108

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractPrincipal component analysis is a useful method for analysing data‐matrices. By analysing separate class models, i.e. disjoint principal component modelling as in the SIMCA or FCVPC programs (developed for supervised and unsupervised principal component analysis respectively), the principal component variance/covariance decomposition (class models) may be used to investigate and interpret the data‐structure of separate classes. The potential of comparing the loadings of variables on subsequent eigenvectors in two class models where the same variables have been used will give information for determining how the variance/covariance in the two datasets differ. This information may then be used either to formulate a hypothesis or to select variables which are specific for the different clas

ISSN:0886-9383    

DOI:10.1002/cem.1180020109

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

摘要:

AbstractClassification rules using non‐parametric linear discriminant functions are often developed from training sets that are not linearly separable. In these situations it is a common practice among inexperienced workers to use many different pattern recognition methods and then select the results that look the best. However, this practice will only increase the risk of spurious results. To document this, we recently carried out a series of Monte Carlo simulation studies to assess the level of chance classification for two different classification algorithms. The level of chance classification for a given dichotomy is found to vary with the choice of the non‐parametric linear discriminant function employed. Although previous workers have indicated that the degree of separation in the data due to chance is only a function of the object‐to‐descriptor ratio, the results of this study suggest ot

ISSN:0886-9383    

DOI:10.1002/cem.1180020110

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

ISSN:0886-9383    

DOI:10.1002/cem.1180020111

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY

ISSN:0886-9383    

DOI:10.1002/cem.1180020102

出版商:John Wiley&Sons, Ltd.    

年代:1988

数据来源: WILEY