Subject Index

white noise and pink noise, examples,. 36f exploratory data analysis, 55 partial transparency transform (PTP), example, 56f future, MLPCA, 59 introduc...
0 downloads 0 Views 183KB Size
Subject Index

Downloaded by 66.128.68.174 on October 29, 2015 | http://pubs.acs.org Publication Date (Web): October 7, 2015 | doi: 10.1021/bk-2015-1199.ix002

A Atmospheric aerosol, applying multivariate curve resolution, 125 conceptual framework, 133 air parcel back trajectory, 144f AZ data set, plots, 137f complex models, 147 emission sources, Dulles International Airport, 152 end members, 134 explicit least squares formulations, advantages, 149 mean contributions, 145t multilinear engine, constraints, 145 multiway data, 150 positive matrix factorization (PMF), 139 simulated data for crustal materials, plot, 135f size-composition-time data, 151 source contributions, 143f source profiles, 142f time synchronization model, 150 unmix, 136 unmix-derived source profiles, 138t conclusions, 152 introduction, 130 fine and coarse particles, volume size distribution, 131f mass balance principle, 131 natural physical constraints, 132 Automotive paints, forensic examination, 195 automotive manufacturer, 215 pattern recognition, wavelet coefficients, 216f conclusion, 218 experimental genetic algorithm, 199 library searching, 200 library spectra, 201 method, 197 search prefilters, 199 spectral alignment, 198 wavelets, 198 introduction, 196 library searching, 217 results, 218t results assembly plant, hierarchical cluster analysis, 205f

assembly plant, principal component analysis, 206f assembly plants, 202t Bramalea/Brampton plant, clear coat paint spectra, principal components, 203f Chrysler, development of search prefilters, 201 components, PC plot, 207f Dodge Main plant, clear coat paint spectra, principal components, 204f Newark assembly plant, plot, 212f paint samples, plant group 11, 210f plant group 12, 211f plant group 13, assembly plants, 212t plant group 13, subplants, 213t set samples, validation, 208f St. Louis plant, clear coat paint spectra, principal components, 204f subplants, 209t Toledo plant, clear coat paint spectra, principal components, 205f validation sample, 214t wavelet coefficients, 208f

C Chemical composition, chemometric modeling chemical composition of microalgae, nutrient availability, 315 novel nonlinear predictor surfaces, 319f prediction surface, 317f predictor variable, 318 introduction, 311 chemometrics preliminaries, 314 experimental preliminaries, 312 sea water microalgae species Dunaliella parva, culture, 313f microalgae, growth dynamics cell culture growth, monitoring, 323 conclusions, 329 culture’s growth dynamics, modeling, 327 Dunaliella salina cells, transmission microscopy image, 324f mono-species cultures, size distribution, 326f produced biomass, modeling, 325

341 In 40 Years of Chemometrics – From Bruce Kowalski to the Future; Lavine, et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2015.

Downloaded by 66.128.68.174 on October 29, 2015 | http://pubs.acs.org Publication Date (Web): October 7, 2015 | doi: 10.1021/bk-2015-1199.ix002

map, source 1 increment, 86f map, source 2 increment, 86f map, source 3 increment, 87f map, source 4 increment, 87f mixture analysis, 78 non-detect values, distribution, 72f normalized TEF-scaled profiles, 76f, 81f PCA scores, 77f PRESS plot, 77f profiles, bulk congener, 74f reasons, TEF-scaling, 75 sediment samples, 82f source 1, best match, 84f source 2, best match, 84f source 3, best match, 85f source 4, best match, 85f source contributions, 79f source profiles, 78f spatial interpretation, 85 TEF-scaled profiles, 74f total dioxin TEQ, 71f variance-scaled profiles, 75f X-residuals, 80f

species-specific maximum cell concentration, 328f nutrient competitors, impact, 319 algae cultures, FTIR spectra, 321f binary species mixtures, 322t microalgae species, 319 Chemometrics and physical organic chemist Bruce Kowalski conclusions, 11 extension, PCA, 7 introduction, 1 Bruce Kowalski, last conference, 3f Tucson conference, 2 linear free energy relationships (LFERs), 4 cross validation analysis, 6 Hammett equation, 5 scores, plot, 7f multivariate calibration, 8 ovarian cancer, detection, 9 blind validation samples, projection, 11f OPLS plot, 10f principal commponents analysis (PCA), 4

K D Dioxin sources in sediments background, 66 Baltic Sea surface sediments, study, 68 polychlorinated dibenzo-p-dioxins (PCDD) and furans (PCDF), structures, 66f study aspects, 68 toxic equivalence factors, PCDD, 67t conclusions, 91 discussion correlation, source 3 increments and mercury concentrations, 90f source 1, 88 source 2, 88 source 3, 89 source 4, 90 introduction, 65 methods, data sources, 69 results, 71 data pretreatments, 73 data screening, 71 dioxins, PCA analysis, 76 frequency, 73f HCA dendrogram, 83f interpretation, source, 82

Kowalski’s vision compositional data, 22 domain, 22f multivariate techniques, 23 partial least squares (PLS) model, 24 discriminant analysis, PLS, 25 introduction, 15 mixture surfaces, 16 classical mixture surface, 18f closest-point projection, 20 error checking, 21 higher-dimensional convex geometry, 17 k-dimensional simplex, 18 refined discriminant analysis, 21 summary, 28

M Maximum likelihood principal components (MLPCA), evolution alternating least squares, 46 algorithm, 47f analytical measurements, errors Bruns, Roy, Professor, 34f characterization, 35

342 In 40 Years of Chemometrics – From Bruce Kowalski to the Future; Lavine, et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2015.

Downloaded by 66.128.68.174 on October 29, 2015 | http://pubs.acs.org Publication Date (Web): October 7, 2015 | doi: 10.1021/bk-2015-1199.ix002

covariance surfaces, error correlation, 39f error covariance matrices, 38f multivariate measurement errors, 33 optical spectroscopy, 37 principle, weighted regression, 33f univariate measurement errors, 32 white noise and pink noise, examples, 36f exploratory data analysis, 55 partial transparency transform (PTP), example, 56f future, MLPCA, 59 introduction, 31 measurement error structures, classification, 47 common error, 48f error matrices, pictorial representation, 51f Faber, Doctor Klaas, with Doctor Bruce Kowalski, 50f heteroscedastic independent errors, 49 projection equations, MLPCA, 51t measurement noise, modeling, 58 MLPCA group members, Doctor Bruce Kowalski, 41f Seattle, pilgrimage, 40 multiway analysis, 56 alternative projection equations, 56 PCA, challenges, 41 singular value decomposition (SVD), 41 subspace modeling, 43 preprocessing data, 53 calibration methods, 54 present, MLPCA, 52 research areas, statistical summary, 53f subspace estimation, 44 Multivariate calibration transfer, essential aspects, 257 calibration modeling, 260 instrument comparison, 261 standardization methods, 261 instrumentation issues spectrophotometers, types, 258 mathematical aspects, calibration transfer, 262 instrument correction, 263 virtual instrument standardization, 263 practices, calibration transfer, 259 test sets of transfer samples, results bias one-sample t-test, 264 bias two-sample t-test, 265

correlation coefficients between parent and child instruments, 266 limit test, 269 r to z’ transformation, 268t slope adjustments, 264 t distribution, critical values, 268t uncertainty, formal statistical methods, 273 Bland-Altman plot, 276 conclusions, 280 difference plotted, sample for instrument A and B, 279f illustration, analytical data, 276t line of equality, correlation between methods, 278f perfect line of equality, data points, 277f standard uncertainty, 274 variation between instruments, global or robust models indicator variables, use, 272 local methods, 272 models over time, augmentation, 271 sample selection, 272 spectral data transformation, 272 Multivariate curve resolution (MCR) constraints, MCR-ALS, 97 closure, 101 correlation, 102 description, 99 examples, 100f hard-modeling, 102 known pure spectra, 101 local rank, 101 model constraints, 102 non-negativity, 100 species, correspondence, 101 unimodality, 100 error propagation, 113 error levels, results, 114f Monte Carlo simulations, error level results, 114f noise propagation, effect, 115f uncertainities, 116 history, 95 iterative target factor analysis (ITTFA) method, 96 Kowalski, Bruce and MCR, 124 with Romà Tauler, 125f MCR solutions, reliability, 109 data matrix, component profiles, 112f rotation ambiguities, calculation of extent, 110 multiset and multiway data analysis, 103 implementation, 105f, 107f

343 In 40 Years of Chemometrics – From Bruce Kowalski to the Future; Lavine, et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2015.

Downloaded by 66.128.68.174 on October 29, 2015 | http://pubs.acs.org Publication Date (Web): October 7, 2015 | doi: 10.1021/bk-2015-1199.ix002

profiles of different components, interaction, 107 quadrilinearity constraint, 106 trilinearity constraint, 104 Tucker3 model constraint, implementation, 108f new application domains environmental studies, 118 hyperspectral imaging, 120 low-spatial-resolution HeLa cell images, 122f metabolomics, 116 quadrilinear model constraint, implementation, 120f spectra resolution, 123f three components, MCR-ALS profiles, 121f untargeted LC-MS MCR-ALS strategy, scheme, 119f workflow, schematic representation, 117f

N Net analyte signal (NAS) discussion images, graphical analysis, 235 NAS modeling, 237 RR and PLS plots, corn data, 234f RR and PLS plots, NMR calibration data, 238f RR and PLS plots, NMR data, 233f RR images, temperature data, 236f sample-wise NAS target modeling, 232 tuning parameter, selection, 231 experimental calibration, 229 corn data, 231 nuclear magnetic resonance (NMR) data, 230 pure component spectra, 230f temperature data, 230 global model selection, NAS measurers, 224 possible model vector, 225f introduction, 221 L-shaped curves, tradeoffs, 229 NAS, fundamentals, 222 depiction, N space, 223f ridge regression (RR), 226 projection, 228f

P Protein secondary structure analysis, 299 conclusions, 308 introduction, 300 peptide backbone, 300f materials and methods CD spectra acquisition, 303 data processing, 303 multivariate analysis, UVRR and CD spectra, 304f sample preparation, 301 secondary structure content of proteins, 302f UVRR spectra acquisition, 302 results and discussion composition profiles, effect of preprocessing, 305 data fusion model, 306f fused CD and UVRR spectra, 307f MCR-ALS model, RMSEC, 308t protein secondary structure, 304 root mean square error of calibration (RMSEC), 308f standard deviation, 305f

R Realistically diverse biochemical data, chemometric modeling biomarkers data degeneracy, 287 data preprocessing, 285 database searches, 286 pattern recognition, 287 conclusions, 294 functionality, 291 future, functionality, 292 interactions larger-scale interactions, 290 pattern recognition, 288 introduction, 283 timeline note, 284 other considerations experimental design, 294 secondary metabolites, 293

S Subspace elimination, adaptive regression approach, mathematics, 242 ARSE algorithm, diagram, 245f

344 In 40 Years of Chemometrics – From Bruce Kowalski to the Future; Lavine, et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2015.

Downloaded by 66.128.68.174 on October 29, 2015 | http://pubs.acs.org Publication Date (Web): October 7, 2015 | doi: 10.1021/bk-2015-1199.ix002

calibration set, 244 interest pure component, ratio of analyte, 243f conclusions, 255 experimental algorithm parameters, 246 data set 1, pure component spectra, 246f data set 2, pure component spectra, 247f data sets, 245 software, 245 introduction, 241 results and discussion 1% noise data set, 250f 5% noise data set, 250f data set 1, 247 data set 2, 252 noise added in wavelet space, data set 1 result, 249t noise in wavelength space, data set 1 result, 248t noise in wavelet space, data set 1 result, 248t noiseless data set, histogram of errors, 248f number of variables used, RMSEP as function, 251f predicted vs true Y values, 252f prediction errors with methyl red, histogram, 253f prediction errors with quinaldine red, histogram, 253f uncalibrated interferent in data set 1, wavelet space, 254t uncalibrated interferent in data set 1, wavelet space with added noise, 254t uncalibrated interferent in data set 2, wavelet space, 255t uncalibrated interferent in data set 2, wavelet space with added noise, 255t

W Watershed data, hierarchical classification modeling class labels, structure advantages, 168 algorithms, 163

analytes, variation in the Ohio Valley region, 172f classification metrics, 164 cluster mean convergence, 178f clustering, uncertainty, 175 data, temporal effects, 171 decision trees, 165 exploratory data analysis, 171f Gibbs sampling, 179f hierarchical class structure, clustering, 175f hierarchical taxonomy, 160 imputation, missing data, 169 mixing coefficients, MCMC settling, 177f multi-label classification, 161 sequential multi-label classification, 166 southeastern USA, clustering, 176f surface water data, clusters, 173 terminal node clusters, 180f tree-structured taxonomy, 162f US geological survey, identified watersheds, 181f USGS data, results of clustering, 179 USGS surface water data, modeling, 168 variogram, identification, 174f water data, censored, 170f water study sampling sites, geographical distribution, 169f conclusions, 191 introduction, 159 multi-label hierarchical model, construction classifier, 183 external test set, 189 hierarchical decomposition, 181 identification, 182 Kriged samples, 189 model depths (TD), 186 modeling predictions, error, 191t probabilistic classifier, 184 reserved samples, training error, 188t steps, chemical measurements, 190 test samples location, 186t tree-structured hierarchical model, 184f, 186 unknown sample, tree descending, 185f USGS surface water data, training error, 188t

345 In 40 Years of Chemometrics – From Bruce Kowalski to the Future; Lavine, et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2015.