RESUMO
Two new approaches to multivariate calibration are described that, for the first time, allow information on measurement uncertainties to be included in the calibration process in a statistically meaningful way. The new methods, referred to as maximum likelihood principal components regression (MLPCR) and maximum likelihood latent root regression (MLLRR), are based on principles of maximum likelihood parameter estimation. MLPCR and MLLRR are generalizations of principal components regression (PCR), which has been widely used in chemistry, and latent root regression (LRR), which has been virtually ignored in this field. Both of the new methods are based on decomposition of the calibration data matrix by maximum likelihood principal component analysis (MLPCA), which has been recently described (Wentzell, P. D.; et al. J. Chemom., in press). By using estimates of the measurement error variance, MLPCR and MLLRR are able to extract the optimum amount of information from each measurement and, thereby, exhibit superior performance over conventional multivariate calibration methods such as PCR and partial least-squares regression (PLS) when there is a nonuniform error structure. The new techniques reduce to PCR and LRR when assumptions of uniform noise are valid. Comparisons of MLPCR, MLLRR, PCR, and PLS are carried out using simulated and experimental data sets consisting of three-component mixtures. In all cases of nonuniform errors examined, the predictive ability of the maximum likelihood methods is superior to that of PCR and PLS, with PLS performing somewhat better than PCR. MLLRR generally performed better than MLPCR, but in most cases the improvement was marginal. The differences between PCR and MLPCR are elucidated by examining the multivariate sensitivity of the two methods.
RESUMO
Infrared emission spectroscopy and multivariate calibration are used to provide a method for the quantitative analysis of liquid samples. Differing forms of the data including second derivative and interferogram representation are used for prediction of sample composition, thickness and temperature. Comparisons are made with transmission measurements of the same samples. In some situations emission measurements may be the preferred method of analysis.
RESUMO
Improvements in process control, which determine production efficiency and product quality, are critically dependent upon on-line process analysis. The technology of the required instrumentation will be substantially expanded by advances in sensing devices. In the future, the hardware will consist of sensor arrays and miniaturized instruments fabricated by microlithography and silicon micromachining. Chemometrics will be extensively used in software to provide error detection, selfcalibration, and correction as well as multivariate data analysis for the determination of anticipated and unanticipated species. A number of examples of monolithically fabricated sensors now exist and more will be forthcoming as the new paradigms and new tools are widely adopted. A trend toward not only on-line but even in-product sensors is becoming discernible.
RESUMO
The sequence of a protein can be graphed as a discrete function and a cross-correlation between any two such number sets produces a similarity score. The scores are used to prepare a phylogenetic tree involving hierarchical cluster analysis, non-linear mapping, and minimal spanning routines. Changes are suggested in the sequences of cytochrome c's from Mediterranean fruit fly, locust, and rattlesnake. The method is faster than existing procedures and does not require human intervention at any stage.
Assuntos
Grupo dos Citocromos c/genética , Filogenia , Proteínas/genética , Aminoácidos , Animais , Evolução Biológica , Drosophila/genética , Glucagon/genética , Secretina/genética , Especificidade da EspécieRESUMO
We classified microorganisms from the clinical laboratory by using information provided by the Gram stain and antibiotic sensitivity profiles obtained with the Bauer-Kirby technique. Approximately 4,000 microorganisms, routinely identified and tested for antibiotic sensitivities in a large hospital microbiology laboratory, were used as a data set for several pattern recognition classification methods: K--nearest-neighbor analysis, statistical isolinear multicomponent analysis, Bayesian inference, and linear discriminant analysis. K--nearest-neighbor analysis yielded the highest prospective classification accuracy for gram-negative organisms, 90%. When those organisms displaying an atypical antibiotic resistance pattern were excluded from the data, the gram-negative classification accuracy improved to 95%. These results are inferior to currently accepted biochemical identification methods. Microorganisms with atypical antibiotic resistance patterns are likely to be misidentified and are common enough (17% of our isolates) to limit the feasibility of routine identification of microorganisms from their antibiotic sensitivities.
Assuntos
Antibacterianos/farmacologia , Bactérias/classificação , Infecções Bacterianas/microbiologia , Reconhecimento Automatizado de Padrão , Bactérias/efeitos dos fármacos , Resistência Microbiana a Medicamentos , HumanosRESUMO
Pattern Recognition is becoming established as a general data analysis tool which has widespread applications in chemistry. Whenever something must be learned from objects (elements, compounds, and mixtures) and a chemical/physical theory has not been sufficiently developed, pattern recognition may provide a solution. Materials production problems, screening applications, source identification and structure analysis are important areas of current interest. It is expected that many more areas of application will open up in the years to come. In short, the "educated guess" is being supported by the computer; at least that is our educated guess.